Abstract
The somatostatin analog octreotide was administered to male and female Sprague-Dawley rats by subcutaneous injection for thirteen weeks at 0 (saline control), 0 (placebo control [mannitol and lactic acid; pH 4.2]), 1.25 mg/kg/day and 2.5 mg/kg/day to explore its potential effect on cutaneous vascular morphology. The placebo caused an increase in the incidence of intimal hyperplasia compared to saline controls in female rats; octreotide increased the incidence and severity of intimal hyperplasia in males and females. Intimal hyperplasia consisted of increased numbers of cells located between the endothelial cell layer and the internal elastic lamina. Severity was based on the degree of compromise of the vascular lumen (regardless of vessel size and number), with severely affected vessels having no visible lumen. Intimal hyperplasia in rats treated with octreotide was considered to be an unexpected and adverse finding, given that this compound and other somatostatin analogs have been investigated as reducers of intimal proliferation or restenosis after angioplasty in humans and that no such lesion has been reported in the literature for this class of compound to date. The induction of intimal hyperplasia by the placebo is also a notable finding; this may be because of the low pH of the formulation.
Introduction
Somatostatin, also known as growth hormone inhibiting hormone (GHIH) or somatotropin release-inhibiting hormone (SRIF), is a hormone that inhibits the release of several pituitary, pancreatic, and other hormones as well as gastrointestinal secretions (Ascoli and Segaloff 1996). It and its synthetic analogs have been used for years to treat a variety of neoplastic, endocrine, and gastrointestinal diseases (de Herder and Lamberts 2003; Eriksson and Öberg 1999). This compound class is also known to have immunomodulatory (Karalis et al. 1994) and angiogenic effects (Dasgupta 2004). Studies have shown that administration of somatostatin and its analogs can reduce intimal proliferation in damaged vessels in laboratory animals (Aavik et al. 2002; Schiller et al. 2002; Ulus et al. 1998; Yumi et al. 1997). Yet these compounds have proven to be less effective in reducing intimal proliferation in some clinical trials (Emmanuelsson et al. 1995; von Essen et al. 1997). We report a case where the somatostatin analog octreotide augmented the incidence and severity of intimal hyperplasia in rats and discuss possible reasons for this occurrence.
Materials and Methods
Octreotide 500 μg/mL (GP-Pharm, Barcelona, Spain), placebo (mannitol [45 mg/kg] plus lactic acid [3.4 mg/mL] in aqueous solution adjusted to pH 4.2 with sodium carbonate; GP-Pharm, Barcelona, Spain), or sterile, isotonic saline solution (0.9% NaCl) was administered by subcutaneous injection to one hundred Sprague-Dawley Crl: CD (SD) rats (Charles River, Saint Germain sur l’Arbresle, France) over a thirteen-week period. The subcutaneous route was chosen for administration because this is the route of administration of octreotide in humans. Rats were approximately six weeks of age at the beginning of the study. They were maintained and handled according to international standards of animal health and welfare in an AAALAC-accredited facility. Seven injection sites, identified by tattoo prior to study initiation, were used in rotation for compound administration, so that each site was injected once every seven days. The volume of compound administered was 2.5 or 5 mL/kg/day. Individual dose volumes were calculated weekly based on the body weight of the animal. Local tolerance at injection sites was evaluated daily. After ninety-one days of treatment, animals were euthanatized in equalized group order by carbon dioxide inhalation and exsanguination and necropsied immediately after death. Samples were routinely collected and processed for histology. Only tissue from the injection sites was evaluated histologically. Three sections of each injection site (middle, lateral, and medial) from all animals in all dose groups were evaluated. Sections were cut at 3–5 microns and stained with hematoxylin and eosin. The location of the injection sites is illustrated in Figure 1.
Statistical analysis of the incidence of intimal hyperplasia by injection site was performed using the Fisher’s Exact Test (Conover 1980a). Statistical analysis of the severity of intimal hyperplasia by group was performed using the Kruskal-Wallis test, a nonparametric, two-sided trend test, and the Wilcoxon test (asymptotic) for pairwise comparisons with the saline control group at individual injection sites (Conover 1980b). The software package used to perform these tests was Roelee (Version 2.05I, P. N. Lee Statistics and Computing, Surrey, UK; release date December 2005).
Results
Intimal hyperplasia consisted of an increase in the number of cells located between the endothelial cell layer and the internal elastic lamina (an undulating band that separates the intimal compartment from the medial compartment of arteries and veins), which caused increased thickness of the vascular wall and decreased luminal patency.
Animals in all groups had intimal hyperplasia. The placebo used in this study (mannitol and lactic acid [pH 4.2]) caused an increase in the incidence of intimal hyperplasia compared to saline controls in females, but not in males. This effect was augmented by octreotide (see Table 1).
Minimal to severe intimal hyperplasia (Grades 1–5, respectively) was observed in vessels of the dermis and subcutis. Small to medium-sized arteries and veins were affected. Severity was based on the degree of compromise of the vascular lumen (regardless of vessel size and number), with severely affected vessels having no visible lumen. Hyperplasia affected either part of or the entire circumference of the intima. Most animals were observed to have minimal hyperplasia.
The incidence (absolute and percentage) and average severity of intimal hyperplasia by injection site and by sex are presented in Table 2. Statistical significance of pairwise comparisons to saline and placebo controls is shown next to the absolute incidence or the average severity value, whereas indications of statistical significance for the comparison of average severity grade across all groups (Kruskal-Wallis statistic) and the trends toward increased incidence and severity are presented in separate rows.
With the exception of Injection Sites 3, 5, and 6 in female saline controls, intimal hyperplasia was observed at all sites in all groups. A significant trend toward increased incidence of intimal hyperplasia compared to saline controls was observed for five of seven injection sites in males (
No clear and consistent correlation between injection site location and lesion incidence could be discerned.
Changes in the severity of intimal hyperplasia occurred across dose groups for six of seven injection sites in males and four of seven injection sites in females (Kruskal-Wallis statistic:
As was observed for lesion incidence, no clear and consistent correlation between injection site location and average severity could be discerned. Female saline control animals consistently had less severe average intimal hyperplasia grades than did animals administered the placebo or the test compound. In contrast, the severity of this finding in male saline controls was mostly higher than that of placebo controls, and on one occasion it was comparable to grades in octreotide-treated animals (Injection Site 4). Photomicrographs representing each grade of intimal hyperplasia are presented in Figures 2 –6.
Vascular inflammation/degeneration/necrosis (a vehicle-related change) was observed at all injection sites in placebo controls and octreotide-treated animals without a consistent relationship to dose. Only one saline control animal was similarly affected at only one injection site. Inflammation of the surrounding dermal or hypodermal tissue was rarely seen. Intimal hyperplasia was not consistently associated with inflammation of the vasculature or the surrounding dermal or hypodermal tissue at any injection site.
Affected vessels in the subcutis were frequently near or within areas of fibroplasia, whereas affected vessels in the dermis were most often observed in areas with no adjacent connective tissue lesions.
Discussion
Intimal hyperplasia is defined as an accumulation of cells (generally smooth muscle cells) between the endothelium and the internal elastic lamina of blood vessels. The pathogenesis of the lesion is as follows: endothelial cell denudation and thrombus formation → accumulation of mitogens at site of injury → activation of smooth muscle cells (SMC), which become proliferative and secretory → migration of SMC through breaks in the internal elastic lamina to the intima → SMC proliferation and synthesis of extracellular matrix (Schiller et al. 2002). This process results in decreased lumen diameter and decreased vessel patency.
It is generally accepted that intimal hyperplasia does not occur in the absence of endothelial cell injury. However, experiments have shown that damage to or removal of endothelial cells is not an absolute prerequisite for vascular smooth muscle proliferation. The positioning of a hollow silicone collar around the carotid artery in rabbits induced intimal proliferation of smooth muscle beneath intact endothelium (Booth et al. 1989), which was verified by light microscopy, scanning, and transmission electron microscopy.
Compounds that are known to induce intimal hyperplasia include estrogen- or progesterone-containing oral contraceptives, ergotamine, and methylsergide maleate (fibroblasts are the primary proliferating cell in elastic and large muscular arteries with the latter two agents) in humans, and allylamine and phosphodiesterase inhibitors in rats (Van Fleet et al. 2002). “Novel immunostimulant drugs” have been shown to induce intimal hyperplasia, endothelial cell proliferation, and vasculitis in dogs (Gopinath et al. 1987). Chronic irritation has also been associated with this finding in cases of thrombotic-related occlusion of central venous catheters (Krzywda and Andris 2005) and dysfunction in transjugular intrahepatic portosystemic stent shunts due to the leakage of bile into the bloodstream (Orlando et al. 2004) in humans.
Somatostatin and its analogs are known to influence vascular function (Badway and Blake 2005). Somatostatin has been shown to inhibit endothelial cell nitric oxide synthase (a molecule involved in angiogenesis) activity linked to somatostatin receptor (SSTR) 3 in vivo (Florio et al. 2003). Investigators have compared the effectiveness of a nonselective somatostatin, CH275 (an analog selective for SSTR 1 and 4), and octreotide (an analog selective for SSTR 2 and 5) in eliminating fibrointimal hyperplasia in rats with induced carotid injury (Aavik et al. 2002). Octreotide has been shown to inhibit vascular insulin-like growth factor and to reduce neointimal thickening after balloon injury of the iliofemoral artery in rats in a dose-dependent fashion (Yumi et al. 1997) and to inhibit myointimal thickening in dogs with arterialized vein grafts (Ulus et al. 1998). Administration of an SSTR 2 analog to rabbits with balloon catheterization-induced injury of the thoracic aorta inhibited neointimal thickening with a bell-shaped dose response (Schiller et al. 2002).
Effect of Formulation pH
The placebo (mannitol and lactic acid) and octreotide formulations used in this study were adjusted to pH 4.2 with sodium carbonate. These acidic solutions were injected subcutaneously once per week at each injection site.
Increased production of angiogenic molecules involved in the paracrine regulation of vascular morphology and function occurs under conditions of reduced tissue pH (Shi et al. 2001). Extracellular acidosis occurring locally in foci of inflammation, ischemia and/or hypoxia, and in large tumors, can induce vascular effects (Christou et al. 2005). Vascular endothelial growth factor (VEGF) has been shown to mediate smooth muscle proliferation in vitro and in vivo (Cardús et al. 2006; Concina et al. 2000; Lazarous et al. 1996) and to stimulate smooth muscle cell migration in vivo (Wang and Keiser 1998). It has also been shown to augment the proliferative effect of fibroblast growth factor-2 on smooth muscle cells in denuded rat carotid arteries in vivo (Couper et al. 1997). mRNA production and secretion of VEGF are enhanced in tumor cells at pH levels of 6.9 to 7.1 in vitro (Shi et al. 2001). Increased mRNA expression for VEGF and basic fibroblastic growth factor (bFGF) and increased bFGF secretion have been demonstrated in endothelial cells exposed to acidotic conditions in vitro (d’Archangelo et al. 2000).
The postinjection pH of the subcutis at the injection sites in our study was likely less than neutral because of the low pH of the placebo formulation. This lower pH may have induced conditions similar to those described above, in which increased concentrations of cytokines and growth factors that act as smooth muscle mitogens are found. The consequences of such decreased pH could explain the slight increase in intimal hyperplasia observed in placebo-treated animals compared to saline control animals.
When octreotide was evaluated in laboratory animals for its effects on vascular injury and repair, target vessels were deliberately injured or otherwise manipulated prior to compound administration. In these studies, octreotide either inhibited intimal hyperplasia or had little to no effect (Aavik et al. 2002; Schiller et al. 2002; Ulus et al. 1998; Yumi et al. 1997). In the exploratory toxicity study described herein, vessels at the injection sites were presumed to be healthy prior to compound administration. The incidence and severity of intimal hyperplasia in our study was augmented by octreotide. It is conceivable that the acidity of the octreotide formulation “primed” the subcutaneous vasculature for the development of intimal hyperplasia by increasing the presence of smooth muscle mitogens, and that an additional primary effect of the analog on somatostatin receptors (see below) caused the increased incidence and severity of this lesion compared to saline controls and placebo-treated animals.
Effect of Selectivity for Somatostatin Receptor Subtype in Vasculature
Five receptors for somatostatin exist; analogs act on different subsets of these receptors and therefore may have different vascular effects in the same test system (Dasgupta 2004). The expression of these receptors varies depending on species, anatomic location, and disease state.
Using PCR analysis, frozen tissue homogenates of normal human arteries (aorta, internal mammary) and veins (saphenous) have been shown to express SSTR 1 and lower levels of SSTRs 2 and 4, but not SSTRs 3 and 5 (Curtis et al. 2000). SSTRs 1, 2, and 4 were also expressed in homogenates of arteriosclerotic popliteal arteries, with no concurrent expression of SSTRs 3 and 5 (Curtis et al. 2000). SSTR 2 has been identified in rat carotid arteries using this technique (Bruns et al. 2000).
More precise localization of receptors has been achieved by several investigators using different methodologies. Vascular endothelium in normal, atopic, and psoriatic human skin has been shown to express SSTRs 1–4, with weak expression of SSTR 5 (Hagströmer et al. 2006). Reubi et al. identified SSTR 2 in the SMC layer of peritumoral veins from various human cancer tissue specimens (Reubi, Horisberger et al. 1994) and in human inflammatory bowel disease (Reubi, Mazzucchelli et al. 1994). They indicated that vessels of non-neoplastic human tissues have few somatostatin receptors (Reubi, Horisberger et al. 1994). Curtis, et al. identified SSTR 1 in endothelial cells of normal and arteriosclerotic vessels but were unable to identify SSTR in vascular smooth muscle (Curtis et al. 2000).
In rodents, SSTR 2 has been identified on endothelial cells in normal and injured rat iliac arteries (Chen et al. 1997). Moreover, it has been reported that receptor subtypes 1 and 4 are expressed three- to four-fold more prominently than subtypes 2 and 5 in the rat aortic vascular wall (Aavik et al. 2002). All five subtypes of the somatostatin receptor have been identified in whole cell extracts of rat aortic vascular smooth muscle cells (Lauder et al. 1997) and in rat thoracic aorta smooth muscle cells in vivo (Khare et al. 1999). Up-regulation of the different SSTR subsets occurred in a time-dependent fashion after vascular injury in this species (Khare et al. 1999). In vitro studies indicated that somatostatin inhibits proliferation of smooth muscle cells by activation of receptors similar to those of human SSTR 5 (Lauder et al. 1997).
The methods used by these investigators to identify the target receptors in different subanatomic locations varied—Chen, Curtis, Hagströmer, and Khare used immunohistochemistry, whereas Reubi used quantitative receptor autoradiography. Autoradiography and immunohistochemistry have been reported to have similar sensitivity for the identification of SSTR 2A in cancerous tissue in humans (Korner et al. 2005), with immunohistochemistry resulting in slightly lower incidence but higher resolution of the target receptor. The reason that investigators Hagströmer, Chen, and Curtis did not identify SSTR receptors in vascular smooth muscle cells using immunohistochemistry is unclear.
SSTR 2 has been shown to have two spliced variants—SSTR 2A and SSTR 2B—in rats (Schindler et al. 1998), mice (Reisine et al. 1993; Vanetti et al. 1993), and potentially humans (Cole and Schindler 2000). The rat variants mediate opposite effects on cell proliferation in CHO-K1 cells in vitro (Alderton et al. 1998). Somatostatin and angiopeptin (an SSTR 2/SSTR 5-specific somatostatin analog) cause proliferation in cells expressing the SSTR 2B receptor, whereas they inhibit cell proliferation in cells expressing the SSTR 2A receptor.
Octreotide targets SSTR 2 and 5 (Dasgupta 2004). If the cutaneous and subcutaneous vessels of rat skin have a receptor profile similar to that of the rat iliac artery, then SSTR 2 expression can be expected and octreotide should be able to act on these receptors. Given the identification of the diverging effects of SSTR 2 isoforms, it is tempting to speculate that the intimal hyperplasia observed in this study may be the result of the binding of octreotide to the SSTR 2B isoform on vascular smooth muscle cells in the skin and subcutis.
The identification of SSTR 5 in rat aortic smooth muscle cells (Lauder et al. 1997; Khare et al. 1999) presents the possibility that SSTR 5 may also be present in rat cutaneous vasculature. The results of the current study suggest that subcutaneous injection of octreotide may have a localized effect that depends on the presence of SSTR 2 and/or 5 in the vasculature of the skin and subcutis. The effect of SSTR 5 expression/binding on smooth muscle proliferation in vivo is currently unknown. Minimal SSTR 5 expression has been found in normal rat aorta, and the level of this expression has been shown not to change with denudation injury (Khare et al. 1999).
Effect of Dose
In a study designed to assess the effects of the selective SSTR 2-agonist somatostatin analog PRL-2486 on neointimal thickening and endothelial cell regrowth following balloon catheterization of the thoracic aorta in rabbits, Schiller et al. (2002) found a bell-shaped dose response for the attenuation of neointimal thickening. Inhibition was observed at the dose of 10 μg/kg/day, whereas no effect was observed at the doses of 5 and 20 μg/kg/day. Foegh et al. (1994) observed a similar response pattern for the effect of angiopeptin on 3H-thymidine incorporation and DNA content of balloon-injured abdominal aorta of rabbits at doses of 2, 20, and 200 μg/kg, where the maximal response was observed at 20 μg/kg. In an in vitro experiment in which the ability of somatostatin to inhibit forskolin-stimulated generation of cyclic AMP in cells expressing the 2A and 2B variants of the rat SSTR 2 receptor was evaluated, low doses of the hormone (3 pM–3 nM) dramatically decreased cAMP, whereas doses above 10 nM were less effective (Schindler et al. 1998). These results represent what is known as a hormetic, or nonmonotonic, dose response (Calabrese and Baldwin 2001). In such cases, the effect of compound administration is stimulatory at low doses and inhibitory at high doses, and U-shaped or inverted U-shaped dose response curves can result. This type of response is often observed for hormones and endocrine-active compounds (Welshons et al. 2003).
Scientific literature indicates that somatostatin and its analogs inhibit intimal hyperplasia, whereas the results of the current study indicate an exacerbation of this effect. All the studies presented in the literature reviewed for the drafting of this publication in which effects of somatostatin or its analogs on intimal hyperplasia were reported used microgram quantities of the compound tested. Compounds were administered by intra-peritoneal (Aavik et al. 2002) or intravenous (Schiller et al. 2002) injection, by continuous infusion via osmotic minipump (Aavik et al. 2002; Häyry et al. 1993—pump location not specified; Häyry et al. 1996; Mennander et al. 1993—subcutaneous pump; Schiller et al. 2002—intraperitoneal pump), or by subcutaneous injection (Aavik et al. 2002; Foegh et al. 1994). The highest peptide dose administered in these studies (500 μg/kg/ day, or 0.5 mg/kg/day) was more than ten-fold higher than the recommended human dose for octreotide and lanreotide (Aavik et al. 2002). The current study with octreotide was performed at 0, 1.5, and 2.5 mg/kg/day—doses that are three- to five-fold higher than those reported in the literature. This raises the possibility that a decreased, or at least an equivalent, incidence and severity of intimal hyperplasia compared to saline and placebo controls may have been observed in this study had lower doses of octreotide been administered. The high doses that were administered may have inhibited the expected restraining effect on smooth muscle proliferation, resulting in a stimulation of proliferation.
Aavik et al. (2002) determined that somatostatin, CHR275 (an SSTR 1,4-selective analog), and octreotide were effective in inhibiting intimal hyperplasia in the rat carotid artery only when administered by daily injection as opposed to continuous infusion. They measured plasma levels and half-lives of somatostatin and CHR275 after a single injection of 200 μg/kg (site of injection not specified) and observed peak concentrations of 360 and 280 pmol/L for somatostatin and CHR275, respectively. The half-lives of these compounds were estimated to be twenty minutes and ninety minutes, respectively. (The investigators had no access to an assay system for octreotide.) In comparison, serum levels measured for both compounds after about ten hours of pump infusion were approximately 110 pmol/L. These levels were maintained during the entire seven-day infusion period. No plasma levels were measured for the other studies reviewed for the drafting of this publication, and none were measured in the current study performed with octreotide.
Because Aavik et al. used octreotide in their investigation, we hoped to gain additional insight into our study results by looking at theirs. The results of their study represent a systemic effect (carotid artery effect resulting from intraperitoneal injection), whereas ours represent a local effect (cutaneous vessel effect resulting from subcutaneous injection). We found the presentation of their study results to be ambiguous; the paper mentions both intraperitoneal and subcutaneous injection but does not provide specifics as to when these routes of injection were used. (It is possible that one of the administration routes was mentioned in error.) Consequently, one cannot know with certainty whether the higher plasma levels that they reported were the result of intraperitoneal or subcutaneous injection. Finally, the investigators were unable to obtain plasma levels for octreotide, the compound that we tested. Therefore, any comparisons or inferences that we might try to make between their study results and ours would be purely speculative.
Effect of Anatomic Location of Vessels
Somatostatin is known to mediate opposing effects in blood vessels, depending on vessel type and anatomic location. In the rat, somatostatin caused vasodilation in saphenous arteries but induced vasoconstriction in saphenous veins (Dézsi et al. 1996). It dilated rat pulmonary arteries (somatostatin-14, but not somatostatin-28; Tjen-A-Looi et al. 1992) but constricted arteries in the rat splanchnic bed (Kravetz et al. 1988) and arterioles on the surface of the parietal cortex (Long et al. 1992). Octreotide and lanreotide (also known as angiopeptin) have both been shown to reduce postprandial splanchnic hyperemia by inducing vasoconstriction in the vessels of the splanchnic bed in humans (Mottet et al. 1998). In contrast, somatostatin is an inflammatory (axon reflex) vasodilator in human skin (Anand et al. 1983; Wallengren 1997), and it is reasonable to presume that octreotide and similar analogs have the same effect. A phosphodiesterase (PDE) inhibitor with vasodilatory properties caused “vascular thickening” consisting of smooth muscle hypertrophy and hyperplasia of the media of arteries and veins secondary to necrotizing arteritis in several anatomic locations in rats after one month of compound administration (Westwood et al. 1990). “Intimal thickening,” characterized by increased cellularity and extracellular material between the internal elastic lamina and the endothelium, contributed to vascular thickening after six months of compound administration. This description of intimal thickening is consistent with the findings of intimal hyperplasia in the current study.
Although Westwood et al. do not indicate whether the vasculature of the skin and subcutis was evaluated for intimal hyperplasia, they do specify that clinical evidence of peripheral vasodilatation was observed as a result of administration of the PDE inhibitor. In contrast, no such clinical evidence was reported in the current study with octreotide.
The cause of the differences in the vasoactivity of somatostatin and its analogs in various anatomic locations and vessel types is unknown, but differential expression of SSTRs and/ or differing actions of these compounds on one or more given receptors are possible reasons for the divergent findings. Similarly, different effects of somatostatin and its analogs on smooth muscle cell proliferation (stimulation versus inhibition) in various locations may also be owing to alternate patterns of receptor expression or dissimilar effects on intracellular pathways activated by receptor binding in those anatomic areas. Examples of differing effects on cell proliferation in cell culture have been reviewed by Lu et al. (2001).
Conclusion
The subcutaneous injection of the somatostatin analog octreotide for thirteen weeks augments the incidence and severity of intimal hyperplasia observed in placebo (mannitol and lactic acid)-treated animals. The acidic environment created by local administration of mannitol and lactic acid may be the cause of the intimal hyperplasia observed at the injection sites of placebo-treated animals compared to saline controls. Recently published literature indicates that there are two subtypes of the SSTR receptor targeted by octreotide and that one of these subtypes mediates smooth muscle proliferation. It is possible that this subtype is targeted by octreotide in the skin, leading to the increased incidence of intimal hyperplasia at the injection sites of octreotide-treated animals. It is also possible that the acidic environment induced by the placebo may potentiate the activity of octreotide on its target receptor, leading to an increase in incidence and severity of intimal hyperplasia in octreotide-treated animals.