help button home button Endocrine Society Endocrine Reviews
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Reubi, J. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Reubi, J. C.
Endocrine Reviews 24 (4): 389-427
Copyright © 2003 by The Endocrine Society

Peptide Receptors as Molecular Targets for Cancer Diagnosis and Therapy

Jean Claude Reubi

Division of Cell Biology and Experimental Cancer Research, Institute of Pathology, University of Berne, CH-3010 Berne, Switzerland

Correspondence: Address all correspondence and requests for reprints to: Jean Claude Reubi, M.D., Division of Cell Biology and Experimental Cancer Research, Institute of Pathology, University of Berne, P.O. Box 62, Murtenstrasse 31, CH-3010 Berne, Switzerland. E-mail: reubi{at}pathology unibe.ch


    Abstract
 Top
 Abstract
 I. Introduction
 II. Definitions
 III. Significance of Peptides...
 VII. Outlook
 References
 
During the past decade, proof of the principle that peptide receptors can be used successfully for in vivo targeting of human cancers has been provided. The molecular basis for targeting rests on the in vitro observation that peptide receptors can be expressed in large quantities in certain tumors. The clinical impact is at the diagnostic level: in vivo receptor scintigraphy uses radiolabeled peptides for the localization of tumors and their metastases. It is also at the therapeutic level: peptide receptor radiotherapy of tumors emerges as a serious treatment option. Peptides linked to cytotoxic agents are also considered for therapeutic applications. The use of nonradiolabeled, noncytotoxic peptide analogs for long-term antiproliferative treatment of tumors appears promising for only a few tumor types, whereas the symptomatic treatment of neuroendocrine tumors by somatostatin analogs is clearly successful. The present review summarizes and critically evaluates the in vitro data on peptide and peptide receptor expression in human cancers. These data are considered to be the molecular basis for peptide receptor targeting of tumors. The paradigmatic peptide somatostatin and its receptors are extensively reviewed in the light of in vivo targeting of neuroendocrine tumors. The role of the more recently described targeting peptides vasoactive intestinal peptide, gastrin-releasing peptide, and cholecystokinin/gastrin is discussed. Other emerging and promising peptides and their respective receptors, including neurotensin, substance P, and neuropeptide Y, are introduced. This information relates to established and potential clinical applications in oncology.

I. Introduction
II. Definitions
III. Significance of Peptides and Peptide Receptors in Cancer
A. Overexpressed receptors as molecular targets
B. Peptides and radiopeptides as targeting agents

IV. Critical Evaluation of Tissues and Methods Used for Peptide Receptor Detection in Vitro
A. Tissue type
B. Receptor protein or mRNA
C. Quantification
D. Morphological identification
E. Choice of methods
F. Pitfalls

V. In Vitro Peptide Receptor Expression in Normal Tissues and Tumors
A. Somatostatin receptors
B. Vasoactive intestinal peptide (VIP) receptors
C. Cholecystokinin (CCK) receptors
D. Bombesin/gastrin-releasing peptide (GRP) receptors
E. Neurotensin receptors
F. Other peptide receptors

VI. Clinical Applications
A. General considerations
B. Diagnostic and therapeutic targeting with radiolabeled or cytotoxic peptides
C. Long-term cancer treatment with nonradioactive, noncytotoxic peptides

VII. Outlook


    I. Introduction
 Top
 Abstract
 I. Introduction
 II. Definitions
 III. Significance of Peptides...
 VII. Outlook
 References
 
IT HAS BEEN a challenge for physicians, in particular for oncologists, to identify a simple tool that has the potential to localize and treat human neoplasms at an early stage of development. About 20 yr ago, monoclonal antibodies became very popular as potential magic bullets to be used in cancer (1); however, this fascinating and simple principle turned out to be much more difficult to transpose into reality than expected, mainly because of the excessive molecular mass (~150 kDa) of antibodies (2, 3). It is only in the past few years that adequate drugs based on antibody or antibody fragments have become commercially available for diagnosis and therapy of cancer, in particular of hematological neoplasias (4). About 15 yr ago, an alternative to radiolabeled antibodies appeared in the form of a small (1.5 kDa) radiolabeled peptide, a somatostatin analog, which led to a major breakthrough in this field. On the basis of the discovery that most human neuroendocrine tumors express a high density of somatostatin receptors (5), it has been possible to develop a method for localizing these tumors and their metastases by in vivo somatostatin receptor scintigraphy (6), using iv injection of a radiolabeled somatostatin analog (7). The tumors, after radioligand binding to their receptors and internalization of the ligand-receptor complex, could thus be identified as hot spots on {gamma}-camera scans (Fig. 1Go) (7). This sensitive procedure is superior to all standard diagnostic tools available today for the detection of specific neuroendocrine tumors, such as gastrinomas (8). From a therapeutic point of view, recent pilot studies using high doses of somatostatin analogs radiolabeled with 90Y have shown a reduction or at least a stabilization of the tumor growth (9, 10, 11, 12). Another successful clinical application has been the long-term use of nonradioactive somatostatin analogs as symptomatic treatment of hormone-secreting neuroendocrine tumors. On the basis of the strong inhibitory effect of somatostatin on hormone secretion, it usually results in a remarkable improvement of life quality, predominantly because of normalization of hormone secretion.



View larger version (20K):
[in this window]
[in a new window]
 
FIG. 1. Principle of in vivo peptide receptor targeting of cancer. The radiolabeled peptide (P) is injected iv into the patient and distributed in the whole body. If the patient has a tumor with cancer cells expressing the corresponding peptide receptor (P-R), the radiopeptide will bind to it and internalize with the receptor into the cell (arrows) where the radioactivity will accumulate. Whole body {gamma}-camera scan will detect the radioactivity accumulated in the tumor, whereas the remaining radioactivity in the body will rapidly be cleared through the kidneys.

 
The molecular basis for such clinical applications is the presence of a high density of somatostatin receptors in these tumors. The in vitro detection of somatostatin receptors in human tumor samples, using methods such as in vitro autoradiography, has therefore, in parallel to clinical applications, been of prime interest during the past decade. In vitro receptor data have predicted the outcome of somatostatin receptor scintigraphy (5, 13) or octreotide therapy (14) and have been used to select tumor types suitable for those clinical applications (7).

Although the clinical use of somatostatin has been refined during the past decade, it remained limited to tumor categories that express somatostatin receptors in sufficiently large quantities, i.e., mainly to neuroendocrine tumors. Therefore, it has been of increasing interest to investigate whether receptors for other regulatory peptides are overexpressed in more common human cancers (i.e., in lung, prostate, colon, or pancreatic carcinomas) to apply a strategy similar to that used with somatostatin. This field of investigation, which appears to be a small niche in the very large oncology field, has gained increasing interest in the past decade. The targeting of overexpressed peptide receptors in tumors by small peptides has become a very strong focus of interest for nuclear medicine. Henry Wagner, at the 100-yr anniversary of nuclear medicine, named the peptide approach in nuclear oncology as one of the most promising fields for the next decade (15); gastroenterologists and endocrinologists are also attracted by the concept of peptide receptor targeting (16, 17).

For a better understanding of these clinical applications, it appears therefore timely to review our current knowledge about peptides and peptide receptors, in particular their tissue expression, their role, or potential applications, in cancer pathogenesis, diagnosis, and treatment.


    II. Definitions
 Top
 Abstract
 I. Introduction
 II. Definitions
 III. Significance of Peptides...
 VII. Outlook
 References
 
Peptides are molecules consisting of several amino acids linked together with peptide bonds. The size of peptides can vary from molecules with only two amino acids to as many as 50. In contrast to proteins, they generally do not possess a well-defined three-dimensional (tertiary) structure. Moreover, peptides do not only exist in natural form but also can be designed synthetically as novel molecules. Thus, their actual number is presently very large. This review will be restricted to physiologically occurring peptides and, within this large group, will focus on the so-called regulatory peptides that include the neuropeptides present in the brain, the gut peptide hormones, as well as peptides present in the vasculature (vasoactive peptides) and peptides of the endocrine system. A list of such regulatory peptides with a link to cancer is found in Table 1Go. Particular attention will be given to somatostatin, vasoactive intestinal peptide (VIP), cholecystokinin (CCK), gastrin-releasing peptide (GRP), and neurotensin.


View this table:
[in this window]
[in a new window]
 
TABLE 1. General characteristics of selected peptides

 
In general terms, these regulatory peptides represent a group of different families of molecules known to act on multiple targets in the human body at extremely low concentrations (5). Targets of these peptides are not only the brain and the gastrointestinal tract, but also the endocrine system, the kidneys, the lungs, and the immune, vascular, and peripheral nervous systems. Therefore, regulatory peptides control and modulate the function of almost all key organs and metabolic processes. Their action is mediated through specific membrane-bound receptors; almost all belong to the group of G protein-coupled receptors. They can influence many intracellular effector systems; for instance, the emerging role of peptides in MAPK pathways, known to play an important role in cell proliferation, or in apoptosis, may contribute to the current interest for peptides in cancer research (18, 19). Receptor subtypes with their own ligand specificity and second messenger systems exist for almost all regulatory peptides, thus increasing the diversity of their mode of action (Table 1Go). These peptides may play prominent roles in not only normal conditions but also pathological processes. They may be factors involved in inflammation, but may also play a receptor-mediated role in cancer and cancer progression (Table 1Go).


    III. Significance of Peptides and Peptide Receptors in Cancer
 Top
 Abstract
 I. Introduction
 II. Definitions
 III. Significance of Peptides...
 VII. Outlook
 References
 
A. Overexpressed receptors as molecular targets
One of the main reasons for the increasing interest for peptides and peptide receptors in cancer is the possibility of receptor targeting, because the peptide receptors are often expressed in many primary human cancers. In several instances, it can even be demonstrated that these peptide receptors are overexpressed in cancer, in comparison to their expression in normal tissue adjacent to the neoplasm and/or in its normal tissue of origin. This aspect will be covered as a main topic of this review in Section V. Basically, one may use these receptors as molecular targets in two ways:

1. Binding sites for radioligands.
The receptors are used primarily as binding sites for a peptide analog, little consideration being given to their biological function. Accordingly, this strategy relies primarily on the presence of tumoral receptors able to bind with high affinity the peptide analogs, and not on a receptor-mediated physiological or pathophysiological action of the peptide. Successful clinical applications of this principle are the diagnostic and radiotherapeutic targeting of peptide receptors with radiolabeled peptides. This procedure takes advantage of one important characteristic of many G protein-coupled peptide receptors, namely that they can internalize into the cell together with their ligand (usually agonists) after receptor-ligand interaction at the cell membrane (20, 21, 22, 23, 24, 25). The remarkable paper by Mantyh et al. (21) in the substance P receptor field should be mentioned as an example; they have demonstrated, with a specific antibody to the neurokinin 1 (NK1) receptor, that the NK1 receptor protein, normally confined to the cell membrane of a neuron population in the dorsal horn, was massively internalized after somatosensory stimulation. Furthermore, they used the internalization mechanism to selectively ablate neurons in the dorsal horn that expressed NK1 receptors (26). Substance P, conjugated to a cytotoxic compound, was infused into the spinal cord. The conjugate, after its internalization in neurons expressing the NK1 receptors, killed them. Somatostatin receptors have also been shown to be internalized to various degrees, depending on the receptor subtype involved; for instance, the commercially available targeting agent 111In-DTPA-[D-Phe1]-octreotide (Octreoscan, Mallinckrodt, Inc., St. Louis, MO) was rapidly internalized in a receptor-specific and temperature-dependent manner (22). The internalized receptor-radioligand complex provides an important and useful accumulation of radiotracer in the cell, thus increasing the radioactive signal at the target site (27). This basic strategy can use either peptides linked to cytotoxic drugs (26, 28) or peptides linked to radioactive isotopes (22). In both cases, the internalized ligand may be able to selectively destroy the targeted cell.

2. Targets mediating functional responses.
Alternatively, one can take advantage of the functional receptors as targets to elicit a particular biological response, using unlabeled, nontoxic peptide analogs over a long period of time. The best example of a peptide receptor-mediated biological response is the inhibition of hormone secretion by somatostatin and its analog octreotide (14, 17, 29). In addition, various peptides have been shown in vitro to play an active role in the growth regulation of many types of tumor cells (30) (Table 1Go), suggesting that long-term treatment with adequate peptide analogs may be able to reduce or stop tumor growth in vivo. In a large variety of animal tumor models expressing various peptide receptors, Schally et al. (31, 32) and Moody et al. (30) were indeed able to demonstrate significant tumor growth inhibition or even to stop the growth by use of specific, nonradioactive, and noncytotoxic peptide analogs.

B. Peptides and radiopeptides as targeting agents
The nature of the peptide itself, in particular its molecular structure and behavior, makes it an attractive compound to act as a bullet targeted at the corresponding peptide receptors (Table 2Go).


View this table:
[in this window]
[in a new window]
 
TABLE 2. Regulatory peptides as targeting agents

 
1. Tissue permeability.
As small and usually hydrophilic molecules, peptides are characterized by an excellent permeability that permits an easy and rapid access to the tumor site after systemic injection. There is one exception: peptides will usually not cross a normal blood-brain barrier and will not enter the brain in significant amounts (<0.1% of total peptide injected) after systemic injection. This is a clear advantage when peripheral organs or tumors are the intended targets; because the brain expresses a high density of most of the peptide receptors, brain targeting could be at the origin of numerous central nervous system side effects. However, peptides may be able to penetrate through the blood-brain barrier when the latter is disturbed, as seen in undifferentiated glial tumors such as glioblastomas. The peptide will then be able to reach the tumor site and remain there (33).

Peptides are usually rapidly excreted from the body. This will occur through renal or hepatobiliary excretion or both, depending on the peptide and the type of structural modifications performed on it.

2. Side effects.
Peptides are physiological compounds and, as such, intrinsically nontoxic, as compared with current chemotherapeutic drugs. Side effects, if they occur, may primarily be due to the physiological actions of the peptides (34, 35) and may be expected after administration of pharmacological doses of the nonradioactive compounds, for instance during long-term treatment of tumors. Conversely, side effects at physiological receptor sites are expected to be negligible if radiolabeled peptides are given for in vivo diagnosis or radiotherapy, because very low peptide doses need to be applied for this purpose (7, 36). Furthermore, because peptides usually play a modulatory role in various biological systems, their actions will often be counterbalanced and possibly annihilated by other hormones, growth factors, or neurotransmitters acting in these same systems. Another important characteristic of regulatory peptides is their usual lack of antigenicity, because their size is small. Their analogs are usually not antigenic either.

3. Stability.
Peptides are quite easily synthesized and modified. They withstand the rather harsh conditions for modification or labeling. However, their natural structural conformation makes them extremely sensitive to peptidases; they are rapidly broken down due to cleavage of peptide bonds by several types of peptidases present in most tissues. Thus, metabolically stable analogs must be developed as a prerequisite for successful clinical applications, in particular for long-term treatments. The best example is the development of the somatostatin analogs octreotide, lanreotide, and vapreotide, that have, compared with natural somatostatin, a much prolonged half-life in plasma and tissue and a longer action (Fig. 2Go). This has only been possible through a long and considerable effort of development (31, 37, 38). For many of the other peptides, only limited efforts have been made to develop peptide analogs having an improved stability in the order of hours, although the proteolytic enzymes as well as the precise peptide bonds being cleaved are well known for many of these peptides. For instance, most of the proteolytic enzymes that have been reported to cleave intact neurotensin are known. This allows a more rational design of suitable analogs with stabilized bonds against metabolic deactivation (39). Furthermore, because regulatory peptides and their receptors are physiological entities that have usually been well characterized a long time ago, it is a great advantage to be able to use previous knowledge of synthesis and structure-activity relationship for a corresponding peptide to design, synthesize, and develop novel peptide analogs that may become useful for clinical applications. The design may involve a higher stability, coupling to radioactive isotopes or to toxic moieties such as doxorubicin (28, 40).



View larger version (28K):
[in this window]
[in a new window]
 
FIG. 2. Development of somatostatin analogs for various clinical applications. A, Human somatostatin. B, Octreotide (Sandostatin) for symptomatic long-term treatment of neuroendocrine tumors. C–F, Octreotide linked to various chelators plus radioisotopes. C, 111In-pentetreotide [111In-DTPA-[D-Phe1]-octreotide (Octreoscan)] for tumor scintigraphy. D, 90Y-DOTATOC (90Y-DOTA-[D-Phe1, Tyr3]octreotide) for radiotherapy. E and F, Second-generation somatostatin radiopeptides, 177Lu-DOTATATE and 68Ga-DOTATOC. DOTATATE has a DOTA-[D-Phe1, Tyr3]-octreotide structure, but with a threonine instead of threoninol.

 
4. Radiolabeling.
Small (8–20 amino acids) peptides are usually large enough to provide an adequate attachment site for a chelator molecule that is sufficiently distant from the binding area to prevent a complete loss of the binding affinity of the peptide to the receptor (Fig. 2Go). Chelators such as diethylenetriaminopentaacetic acid (DTPA) or 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) are indispensable molecules to accept certain types of metallic radioisotopes (40). However, those relatively large molecules may affect the binding properties of the compound to which they are attached. Peptidic analogs of regulatory peptides are, due to their size, usually more suitable for chelator attachment than the smaller nonpeptidic analogs.

IV. Critical Evaluation of Tissues and Methods Used for Peptide Receptor Detection in Vitro
In vitro evaluation of peptide receptor expression, in particular of their incidence and density in human tumors and their metastases, is critical for identifying and defining 1) potential peptide receptors that may be of interest for clinical applications, and 2) tumor types that are particularly suitable to be targeted with a given peptide. It has been demonstrated that extensive in vitro information about the receptor incidence and density in a given tumor is required before in vivo investigations can be performed in humans. However, the various in vitro methods available for peptide receptor identification and for the prediction of suitable targets can provide variable information (Table 3Go). Therefore, it is important to critically review and evaluate those methods to understand their respective value.


View this table:
[in this window]
[in a new window]
 
TABLE 3. In vitro evaluation of peptide receptor expression

 
A. Tissue type
To predict the in vivo peptide receptor status of a tumor as closely as possible, the in vitro method used should evaluate the peptide receptors in primary human tumors and their metastases. Such tissues are usually available after surgical resection. There is increasing evidence to suggest that it is not sufficient to have identified receptors in tumor cell lines or in tumors grown in animals because there might be profound differences in receptor expression, not only between species but also between cultured tumor cells or tumor cells growing in situ (41, 42, 43, 44). Peptide receptor expression in cell lines does not necessarily reflect the peptide receptor situation in the corresponding human primary tumors (41, 45, 46). Igarashi et al. (44) recently showed not only that species differences exist for the VIP receptor subtype 1 (VPAC1) pharmacophore but, more importantly, that human VPAC1 expressed in Chinese hamster ovary cells differed markedly from the native human VPAC1. It is therefore essential not only to be aware of possible species differences but also to investigate systems expressing the native receptor, in particular when searching for new peptide analogs for clinical use.

B. Receptor protein or mRNA
Because it is the receptor protein that is ultimately targeted in vivo, it should also be the protein that is investigated in vitro, rather than the receptor mRNA. In human tumors, the mRNA and the protein levels for selected peptide receptors may differ (47, 48). Therefore, it is particularly relevant, as first priority, to gain information about receptor protein expression. Receptor binding methodologies (binding assay; receptor autoradiography) or receptor immunohistochemistry provides such basic information on receptor protein. The first of these methods is particularly suitable because it represents the closest correlation to in vivo binding as shown by in vivo receptor scintigraphy; all clinical applications presently available are based on binding of a peptide analog to the receptor binding site. Receptor immunohistochemistry, on the other hand, has an excellent cellular resolution, but may, however, identify an epitope that is not identical with the binding site and may therefore not tell much about the binding capability of a tumoral receptor. The information about the mRNA abundance is important when the methods measuring receptor proteins are unable to give enough clues, for instance about receptor subtypes.

C. Quantification
It is necessary to evaluate not only the incidence of a peptide receptor in a given tumor type, but also its density. In particular, radiotherapeutic implications will rely mainly on tumors with a high receptor density (11, 12, 49). Thus, an in vitro method that quantifies the amount of peptide receptors is preferable. Because peptide receptors may exist in the form of multiple subtypes, it is important to use an in vitro method that is specific enough to detect these subtypes.

D. Morphological identification
To assess the tumor to background ratio of peptide receptors present in a given tissue, the receptors should be evaluated not only in tumor material but also in normal tissue, preferably tissue adjacent to the tumor and containing the tissue of origin of these tumors. We must remember that samples of human tissues, for instance surgically resected tumors, are morphologically complex: the tumor is often intermingled with a variety of nonneoplastic tissue, such as leukocytes, vessels (including newly formed vessels), reactive components (fibrosis, necrosis), as well as adjacent normal host tissue elements, such as epithelia (e.g., in the mucosa of the gastrointestinal tract), smooth muscle, nerve, immune cells, or stroma that may in many instances express the corresponding peptide receptors. These facts imply that it is mandatory to evaluate the receptor protein of human tissue samples with morphological methods to identify precisely which tissue elements express the receptors of interest. Those elements cannot be identified in tissue homogenates.

E. Choice of methods
One of the methods that fulfills most of the above-mentioned requirements is in vitro receptor autoradiography (Table 3Go). It localizes and quantitates the peptide receptor. It identifies the receptor protein through its binding site. Pharmacological experiments using subtype-selective analogs allow the gross identification of peptide receptor subtypes by their rank order of potencies in displacement experiments (50, 51, 52, 53). The method has a high sensitivity when 125I-labeled peptides are used. One of the disadvantages of the method is, however, its limited cellular resolution. An accumulation of several cells of the same type is necessary to generate a radioactive signal that can then be attributed to those cells expressing the peptide receptor. Such an accumulation of similar cells is usually not a problem in a solid tumor, but it may be difficult to find in a normal tissue with a complex histological pattern.

Binding sites can also be determined with the less time-consuming in vitro binding assays. In contrast to receptor autoradiography, these are performed on homogenates, lack the morphological correlates that are necessary in this type of investigation, and are therefore less adequate for the study of surgically resected material. In human tumor samples, receptor quantification with binding methods will usually express the receptor density per milligram of tissue or protein, in contrast to cell cultures in which the number of receptors can be expressed per cell. Receptor immunohistochemistry identifies the receptor protein with a much better cellular resolution than autoradiography. However, the proteins cannot be reliably quantified, and the assay may recognize proteins unrelated to the receptor binding site (54). Moreover, specific receptor antibodies are not available for each of the peptide receptor subtypes under investigation.

In situ hybridization, Northern blots, RNase protection assays, RT-PCR, and real-time-PCR all identify mRNA. PCR and RNase protection assays are methods with a very high sensitivity, but without morphological correlates. In situ hybridization, on the other hand, is a morphological method, however, with a lower sensitivity than RT-PCR. Recently, first reports on measurement of sst2 receptor by real-time PCR indicate (55) that mRNA could be quantified precisely. It should be remembered that RT-PCR identifies extremely low message levels that may not translate into the expression of functional levels of receptors.

F. Pitfalls
In all investigations of human material, one deals with considerable individual variations in the measured parameters from one individual to the other, not only in tumors but also in normal tissues. Such a variability is usually not observed in studies using laboratory animals or tumor cell lines cultured in vitro. Awareness of the potential problems associated with human tissue sampling is indispensable (Table 3Go). It is difficult to standardize tissue sampling and processing: the sample size (biopsy, surgical resection), the delay after sampling (surgical, postmortem), and the sample processing (fresh, frozen, fixed; method of fixation) are variable parameters (Table 3Go). Inadequate storage (resected material not frozen immediately) and/or processing (thawing of tissue) may easily lead to false-negative receptor data. A resected tissue sample often displays multiple concomitant pathological changes (neoplasia, inflammation, adaptive changes) in a heterogeneous topographical distribution; furthermore, adjacent normal tissues may also reveal pathological features. Such constituents may be receptor-positive. These facts further emphasize the need for a morphological evaluation of the receptor distribution, if possible in cooperation with an experienced pathologist.

V. In Vitro Peptide Receptor Expression in Normal Tissues and Tumors
This section summarizes the peptide receptor characteristics of human tissues, in particular tumors, based on current in vitro information. Subsections have been dedicated to receptors for somatostatin, VIP, CCK, GRP, and neurotensin, because those receptors have been extensively studied and because all of them have been targeted in vivo in patients. Another subsection will deal with additional peptide receptors, for which less information is available or which recently began to be evaluated.

A. Somatostatin receptors
1. General background.
Somatostatin consists of a family of a 14-amino-acid (somatostatin-14) and a 28-amino-acid (somatostatin-28) peptide (Table 1Go and Fig. 2Go). It appears in several organ systems, such as the central nervous system, the hypothalamopituitary system, the gastrointestinal tract, the exocrine and endocrine pancreas, and the immune system. It inhibits a wide spectrum of physiological functions, including peptide hormone secretion. In these different organ systems, somatostatin can be considered to be a neurotransmitter, a neurohormone, or a local hormone acting via autocrine or paracrine mechanisms (56). Moreover, somatostatin plays a role in cancer: in many animal tumor models and cultured tumor cell lines, somatostatin and somatostatin analogs inhibit tumor growth (30, 31).

Somatostatin and octreotide actions are mediated by specific, high-affinity somatostatin receptors located on the plasma membrane of the target cells. To date, five human somatostatin receptor subtypes (sst1, sst2, sst3, sst4, and sst5) have been cloned and partially characterized (57, 58) (Table 1Go). They have distinct, often overlapping patterns of expression in human organs. These subtypes belong to a superfamily of G protein-coupled receptors that can functionally couple to various intracellular effector systems. One of the most widely studied systems is the adenylylcyclase-cAMP-protein kinase A pathway that can be inhibited by somatostatin in numerous cell types (56); furthermore, the modulation of potassium channels by somatostatin has been extensively documented (56). Other relevant signaling pathways regulated by somatostatin include the somatostatin-induced stimulation of phospholipase A2 (59) and the activation of phosphotyrosine phosphatases (60). Particular interest has been devoted recently to the inhibitory action of somatostatin through MAPK pathways, probably an important facet of somatostatin signaling (19). Because MAPK activation plays an important role in cell proliferation (61), inhibition of MAPK pathways is likely to contribute to the antiproliferative effect of somatostatin. The role of somatostatin in stimulating apoptotic mechanisms in sst2- or sst3-expressing cells (62, 63) is another notable antiproliferative mechanism. Remarkably, somatostatin receptors have recently been shown to form homo- and heterodimers (64, 65, 66) and to physically interact with a class of proteins displaying anchoring and scaffolding functions (67, 68, 69). Pharmacological studies revealed that all five human subtypes bind somatostatin-14 and somatostatin-28 with a high affinity. However, there are differences in the binding affinities of the structural analogs of somatostatin; for instance, octreotide is bound with high affinity by the sst2 and sst5 receptor subtypes and with a moderate affinity by sst3, but not by subtypes sst1 and sst4 (58). There are also differences in the cell trafficking and internalization capabilities of the different receptor subtypes (20, 56); upon ligand binding, sst3 and sst2 internalize much better than sst1 (20). sst5 Not only internalizes after ligand binding but can additionally trigger a massive recruitment of sst5 receptors from intracellular stores to the membrane (70). Presently, several groups in academia and in the pharmaceutical industry are searching for somatostatin analogs with binding profiles selective for specific subtypes.

2. In vitro detection of somatostatin receptor in normal human tissues.
The information on somatostatin receptor and receptor subtype distribution in normal human tissues is still incomplete, mainly because of limited access to normal human tissues. Unfortunately, it is not possible to simply extrapolate animal data to humans because of species differences (71, 72). Receptor binding studies, mRNA determination, and/or receptor immunohistochemistry have identified somatostatin receptors in human brain (for review, see Ref.73) as well as in numerous peripheral tissues, including pituitary, pancreas, gut, thyroid, adrenal, kidney, and the immune system; a complex pattern of somatostatin receptor subtype expression has been observed, including coexpression of multiple subtypes in a tissue-specific pattern (74, 75, 76). The subtype most frequently expressed is usually sst2A, as shown in recent immunohistochemical and receptor autoradiographical studies using subtype-selective antibodies and somatostatin analogs; abundant sst2A is found in pancreatic islets (72), in specific regions of the human brain (77), and in the peripheral nervous system (plexus myentericus and submucosus) (54). It shows up in the immune system, i.e., in the germinal centers of lymphoid follicles and in human peripheral blood lymphocytes (54, 78). sst2 is also present in the human adrenal gland (79) as well as in the kidneys (80, 81). The precise localization of the other ssts in human tissues is not yet fully established. sst3 and sst5 have been identified in T lymphocytes (82, 83, 84). The human placenta (85) as well as the fetal and adult lung display predominantly sst4 (86, 87).

3. In vitro detection of somatostatin receptors in human tumors.
Somatostatin receptors are not only expressed in physiological conditions. Indeed, a wide variety of human tumors express somatostatin receptors, which can be detected in vitro (Table 4Go).


View this table:
[in this window]
[in a new window]
 
TABLE 4. Peptide receptor expression at the protein level in human tumors

 
   a. Receptor binding.
Early studies using in vitro receptor autoradiography have provided the identification of human tumors expressing somatostatin receptors, but without giving, at that time, information about subtypes, which had not yet been discovered. A very high incidence and often a high density of somatostatin receptors have been found in neuroendocrine tumors, in particular in GH-secreting pituitary adenomas (88, 89, 90) and gastroenteropancreatic tumors (91, 92), pheochromocytomas (79, 93), neuroblastomas (94, 95), and to a lesser extent in medullary thyroid cancers (96) and small cell lung cancers (97). Nonneuroendocrine tumors including brain tumors such as meningiomas, medulloblastomas, or gliomas (98, 99, 100), but also breast carcinomas (101, 102), lymphomas (103), renal cell cancers (104), mesenchymal tumors (105), prostatic (106, 107, 108), ovarian (109, 110), gastric (45, 111), hepatocellular (112, 113) and nasopharyngeal carcinomas (114) were also shown to express somatostatin receptors. The main classes of tumors that express somatostatin receptors are listed in Table 4Go. Various radioligands were used in binding studies, either natural somatostatin, e.g., 125I-analogs of somatostatin-14 or somatostatin-28 as universal radioligands, or synthetic, small-sized analogs, such as 125I-[Tyr3]octreotide, 125I-MK-678, or 125I-RC-160, which label only selected somatostatin receptor subtypes (115). In general, these studies revealed that somatostatin receptor expression was highly variable from one individual to another and from one tumor type to another. Whereas some tumors are characterized by a high density of receptors, such as meningiomas or medulloblastomas, others, such as lymphomas, have a much lower density. Some tumors have a rather homogeneous somatostatin receptor distribution, e.g., most neuroendocrine tumors, in particular gastroenteropancreatic tumors, as shown by a somatostatin receptor-positive carcinoid in Fig. 3Go. Other tumors, such as breast carcinomas, are characterized, however, by a highly heterogeneous somatostatin receptor distribution, with regions of high density next to regions lacking the receptor (101), a pattern that reflects the marked polyclonality of this type of tumor. The determination of a precise value of receptor density is therefore hardly possible in such tumors, and only of relative significance. Receptor homogeneity is very important with regard to potential targeting of these somatostatin receptors for diagnosis or therapy.



View larger version (115K):
[in this window]
[in a new window]
 
FIG. 3. In vitro detection of somatostatin receptors in carcinoids. A–C, Receptor autoradiographical analysis. A, Hematoxylin-eosin-stained section. Scale bar, 1 mm. B, Autoradiogram showing a high density of somatostatin receptors in the whole tumor (total binding of 125I-[Leu8, D-Trp22, Tyr25]-somatostatin-28). C, Autoradiogram showing nonspecific binding (in presence of 10-6 M somatostatin-28). D–F, In situ hybridization of sst2 mRNA in the same carcinoid. D, Hematoxylin-eosin-stained section. Scale bar, 1 mm. E, Autoradiogram showing sst2 mRNA in the tumor by use of a 33P-labeled probe. F, Control section, in presence of excess of unlabeled probe.

 
As stated in Section IV, a crucial problem of in vitro analysis of somatostatin receptors in tumor samples is the incidence of false-positive results due to amalgamation of the tumor samples with nonneoplastic, somatostatin receptor-expressing tissues. For instance, most of the colorectal cancers minimally express octreotide binding sites (116, 117), whereas the vessels located around the tumors (peritumoral vessels) have a high density of such sites (116). In glial tumors, even morphological methods such as receptor autoradiography may not be able to adequately distinguish between somatostatin receptors present in tumor cells and those present in residual nerve fibers (99, 118). These glial tumors are known, indeed, to heavily infiltrate the normal brain, which itself strongly expresses somatostatin receptors. The high proportion of glial tumor samples contaminated by somatostatin receptor-positive nervous tissues makes it difficult to assess precisely the degree of somatostatin receptor expression by these tumors; although both Dutour et al. (119) and Held-Feindt et al. (120) could identify glial tumor cells expressing sst2 with high resolution techniques, such findings were questioned in a more recent study (118).

Most of the receptor autoradiography studies done in the 1980s and early 1990s still epitomize the basic, still valid information on somatostatin receptor expression in tumors although somatostatin receptor subtypes could not precisely be identified at that time. Analysis of mRNA, immunohistochemistry with selective antibodies, and autoradiography with subtype-selective ligands have recently been introduced and have been able in most instances to confirm the previous data and extend them by determining the subtypes involved. However, by far not in all tumor types were these methods able to bring clarity in terms of the sst subtype protein involved.

   b. Receptor mRNA.
Human tumors often express multiple somatostatin receptor subtype mRNAs, as reported first in pituitary adenomas (121, 122, 123, 124, 125) and gastroenteropancreatic tumors (125, 126, 127). A carcinoid with abundant sst2 mRNA is shown in Fig. 3Go. In the past few years, a profusion of papers have appeared identifying mRNA for the various somatostatin receptor subtypes in a large variety of other human cancers (79, 108, 109, 119, 125, 128, 129, 130, 131, 132, 133, 134, 135, 136, 137, 138), confirming and extending the results of previous binding studies. In many of the RT-PCR-based investigations, the incidence of the various receptor mRNAs in tumors appears, however, to be higher than that detected by receptor binding or immunohistochemical studies (41, 48, 101, 116, 139); moreover, mRNAs for several somatostatin receptor subtypes appear to be frequently expressed concomitantly in individual tumors. It is presently not clear whether 1) these RT-PCR data reflect an overestimation of the real contribution of the various somatostatin receptor mRNAs due to the outstanding sensitivity of the method, 2) these mRNAs originate in part from nontumoral adjacent tissues, or 3) these mRNAs are not always translated into significant amounts of the respective receptor subtype proteins. Thus, one may caution against an overestimation of mRNA data obtained by ultrasensitive methods such as RT-PCR. We should not rely exclusively on mRNA determinations to assess the tumoral receptor status; the main target for the current clinical applications of somatostatin ligands is the receptor protein located on the cell membrane, not the mRNA.

   c. Receptor immunohistochemistry.
An emerging new technique to detect somatostatin receptor subtypes is immunohistochemistry, which has the advantage of a high cellular resolution. The results, however, depend clearly on the quality, selectivity, and specificity of the antibodies, several of which directed against somatostatin receptors are currently available. Even highly specific antibodies for sst2A, such as R2-88, can weakly cross-react with unrelated proteins (54). Up to now, only a few, carefully controlled immunohistochemical studies have been performed in cancer, primarily with sst2 antibodies. A high density of sst2A was found in neuroblastomas (140), medulloblastomas (137, 140), paragangliomas (140), and small cell lung cancers (140) as well as in meningiomas (141) and breast cancers (142, 143). Most neuroendocrine lung and gastroenteropancreatic tumors were shown to have preferentially membrane-bound sst2 with immunohistochemical methods (144, 145, 146, 147, 148). An example of an sst2A-expressing carcinoid with strong membrane-bound receptor localization using R2-88 is shown in Fig. 4Go. There have been few reports investigating other somatostatin receptor subtypes with immunohistochemistry, such as sst1, sst3, and sst5 in gastroenteropancreatic tumors (147, 149) and sst3 in breast cancers (142). The observation of an intracellular location of some of these receptors (147, 149) is intriguing and not fully understood. It is, however, worth noticing that, both in rat brain and human tumors, a correlation between the local presence of endogenous somatostatin and an increased internalization of sst2 receptors could be observed (140, 150).



View larger version (151K):
[in this window]
[in a new window]
 
FIG. 4. sst2A In a carcinoid tumor. Immunohistochemical detection of membrane-bound sst2A receptors with R2-88 antibody. The strong red-brownishmembrane-bound immunostaining on each of the tumor cells reflects the abundance of somatostatin receptors of the sst2A subtype in this ileal carcinoid. (R2-88 antibody was generously provided by Dr. A. Schonbrunn, Houston.) Scale bar, 0.1 mm.

 
Analysis of somatostatin receptors in tumors by immunohistochemistry may become an additional, useful parameter for the clinician to evaluate the biology of a tumor and choose therapeutic options, in particular because standard formalin-fixed material is sufficient for this type of investigation (145).

   d. Binding studies with subtype-selective analogs.
The recent development of somatostatin receptor subtype-selective analogs, both as peptides and nonpeptides (83, 151, 152, 153), is an important advance permitting 1) evaluation of the distribution of various receptor subtype proteins in tissue, 2) determination of the specific biological effects mediated by the various subtypes, and 3) the design of new drugs for specific therapeutic strategies. Some of these analogs were already used to refine somatostatin receptor binding studies to detect receptor subtypes in tissues (154, 155). Moreover, in a study using receptor autoradiography with five different subtype-selective analogs, we evaluated somatostatin receptor subtypes expressed in cancers (50, 156); these data suggest that in many somatostatin receptor-positive tumors there is a predominance of the proteins for one or two somatostatin receptors. A preponderance of sst2 binding sites is seen in the majority of neuroblastomas, medulloblastomas, breast cancers, meningiomas, paragangliomas, renal cell carcinomas, lymphomas, hepatocellular carcinomas, and small cell lung cancers (50). Conversely, sst1 is frequent in prostate cancers and in many sarcomas, whereas sst3 occurs frequently in inactive pituitary adenomas (50, 83). A larger subtype variability with several ssts expressed concomitantly is seen, among others, in GH-producing pituitary adenomas (especially sst2 and sst5), pheochromocytomas, hormone-producing gastroenteropancreatic tumors, and gastric cancers (50, 156, 157). Interestingly, sst4 is not often expressed in the human cancers tested.

4. Somatostatin receptors in peritumoral vessels.
Recently, the peritumoral vascular system of the host has emerged as a possible target of somatostatin action in tumor development. In a series of human colonic carcinomas, a high density of vascular somatostatin receptors was observed in vessels in the immediate vicinity of the tumors; the receptor density decreased continuously with increased distance of the vessels from the carcinomas, suggesting a local phenomenon related to the presence of the tumor (116). The presence of vascular somatostatin receptors seemed to be independent of the presence or absence of somatostatin receptors in the tumor itself. More recently, a study including a large number of different types of human neoplasms has suggested that the expression of somatostatin receptors in peritumoral veins is a general phenomenon (158). For instance, all medullary thyroid carcinomas, colonic, and gastric cancers express somatostatin receptors in peritumoral veins; a majority of parathyroid adenomas, renal cell cancers, melanomas, sarcomas, breast cancers, and prostate cancers have somatostatin receptors in peritumoral veins, whereas gastroenteropancreatic tumors or ovarian cancers rarely do. Recent studies demonstrated that angiogenic vessels as well as peritumoral vessels expressed predominantly sst2 (119, 159). In some tumors, such as melanomas, the somatostatin receptors are expressed not only in peritumoral but also in intratumoral veins. This may be the reason why it is possible to successfully visualize this type of tumor in vivo with Octreoscan, although melanoma cells do not express significant levels of somatostatin receptors.

The function of somatostatin in the peritumoral vasculature, mediated by a high density of somatostatin receptors in the smooth muscle cells and possibly in the endothelium (158, 159), may be primarily vasoconstrictive, as shown in particular in the gut (160). Therefore, an increased somatostatin receptor density may allow a strong and rapid local vasoconstriction, possibly resulting in local hypoxia and necrosis of the tumor, or a more prolonged vasoconstriction, directed against metastatic tumor dissemination. Whether this mechanism is responsible for the occasional clinical observation of a decrease in tumor size during long-term octreotide therapy in some patients (161) is unknown. Despite a very broad interest in tumor angiogenesis in general, progress in understanding the role of peritumoral somatostatin receptors has been very slow (162). In neoplasms, somatostatin may act locally on tumor growth through two different mechanisms dependent on local somatostatin receptor expression: through direct action on tumor cells or through action on peritumoral vessels, which may alter the dynamics of the tumoral blood circulation and/or inhibit angiogenesis (162).

5. Somatostatin receptors in nonneoplastic diseases.
There is strong evidence that selected nontumoral lesions may also express somatostatin receptors. For instance, active granulomas in sarcoidosis express somatostatin receptors on the epithelioid cells (163). Inactive or successfully treated fibrosing granulomas devoid of epithelioid cells lack somatostatin receptors. Inflamed joints in active rheumatoid arthritis express somatostatin receptors, preferentially located in the proliferating synovial vessels (164). Furthermore, inflammatory bowel disease is characterized by an overexpression of somatostatin receptors in the vascular system (165) of the altered parts of the gastrointestinal tract. The expression of somatostatin receptors is therefore not specific for tumoral pathologies.

B. Vasoactive intestinal peptide (VIP) receptors
1. VIP/pituitary adenylate cyclase activating peptide (PACAP) and their receptor subtypes.
VIP is a 28-amino-acid-long neuropeptide isolated from the small intestine. It is a member of the group of secretin-like peptides (166). Together with PACAP, a structurally similar, 27- or 38-amino-acid-long peptide, it is one of the important neurotransmitters in the gut. VIP and PACAP both play a neuromodulatory role in the central nervous system, at both the neuronal and glial levels (167). Furthermore, extensive immunomodulatory properties have been reported for these peptides (168, 169, 170, 171). Their actions are mediated by specific G protein-coupled receptors that can be internalized after ligand binding (25). One of the most prominent signaling pathways of VIP/PACAP is the stimulation of adenylate cyclase activity, as seen impressively in nonfunctioning pituitary adenomas (172). In the last few years, molecular biology has provided evidence for the existence of several receptor subtypes within the VIP/PACAP family (173, 174). There are two VIP receptors, VPAC1 and VPAC2, both with high affinity for VIP and PACAP (Table 1Go). They can be distinguished pharmacologically by the VPAC1-selective analog [Lys15,Arg16,Leu27]VIP(1, 2, 3, 4, 5, 6, 7)/GRF(8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27) (KRL-VIP/GRF) and the VPAC2-selective RO 25-1553 (175, 176). There is at least one PACAP receptor, named PAC1, that is characterized by high affinity for PACAP but by a low affinity for VIP (173, 174, 177); recently, several PAC1 splice variants with distinct pharmacological behaviors have, however, been identified (178, 179).

2. VIP/PACAP receptors in normal human tissues.
VIP/PACAP receptors are found not only in the brain (167), but ubiquitously in the majority of the human epithelial tissues (180). In most of these tissues, the VIP/PACAP receptor subtype preferentially expressed is the VPAC1 receptor, for instance in hepatocytes, gastrointestinal mucosa, lobules and ducts of the breast, thyroid follicles, prostatic glands, urothelium of bladder and ureter, and acini of the lung and pancreatic ducts (52, 180, 181). Some other tissues, however, predominantly express the PAC1 receptor, e.g., the adrenal medulla, several brain areas, and the pituitary (52). Conversely, smooth muscle in various locations preferentially expresses VPAC2 receptors, as documented by 125I-VIP binding displaced by nanomolar concentrations of the VPAC2-selective RO 25-1553, but not of the VPAC1-selective analog KRL-VIP/GRF (52). Such VPAC2 receptors in smooth muscles are present in locations as different as the gastrointestinal tract (stomach) or the seminal vesicle. Furthermore, blood vessels, arteries more than veins, express VIP receptors of the VPAC2 subtype located primarily in the smooth muscle layers. Moreover, the wall, i.e., most likely the smooth muscle, of the uterus and also of the prostate are primarily VPAC2 receptor-expressing tissues. The peripheral nervous system, for instance the myenteric plexus in the colon wall, shows a predominance of PAC1 receptors, whereas the Cajal cells may express VPAC2 (182). A great majority of the human solid lymphoid tissues, including spleen, thymus, lymph nodes, and Peyer’s patches express VIP/PACAP receptors at high density (169, 183). This presence of VIP/PACAP receptors in most normal human tissues points to multiple and complex biological actions of VIP/PACAP in the human body.

3. VIP/PACAP receptors in human tumors and their metastases.
Human tumors derived from normal tissues expressing VIP/PACAP receptors frequently also express the same receptors. While the great majority of tumors analyzed earlier for VIP/PACAP receptors were actually tumor cell lines (Refs. 184 and 185 ; for review, see Ref.174), very few authors have investigated primary human cancers. A few years ago, it was shown that most primary human tumors express VIP/PACAP receptors at high incidence (105, 186). In these studies, 125I-VIP binding displaced by VIP or PACAP was investigated. More recently, a follow-up study tried to discriminate the subtype expressed by these tumors using receptor subtype-selective analogs (52).Tumors expressing VPAC1 receptors include the most frequently occurring malignant epithelial neoplasms, such as cancers of the lung, stomach, colon, rectum, breast, prostate, pancreatic ducts, liver, and urinary bladder (52) (Table 4Go). In contrast to this ubiquitous expression of VPAC1 receptors in most human tumors, a predominance of VPAC2 receptors was found in only few tumors. The only example of a consistent and predominant VPAC2 receptor expression, among the tumors tested, is that of the benign smooth muscle tumors, the leiomyomas (Table 4Go). They exhibit a strong 125I-VIP binding displaced with nanomolar concentrations of RO 25-1553 but not of KRL-VIP/GRF; they can also be labeled directly by 125I-RO 25-1553 (52). In contrast, several different human tumor types express predominantly PAC1 receptors, in particular tumors originating from the neuronal and endocrine systems; this includes glial tumors (astrocytomas, glioblastomas, oligodendrogliomas), neuroblastomas, as well as various pituitary adenomas (especially GH-secreting and nonfunctioning adenomas, but not prolactinomas), as described previously by the group of Robberecht (177, 187, 188) and by Oka et al. (189). More recent data have shown that most catecholamine-secreting tumors, including both pheochromocytomas and paragangliomas, appear to express predominantly PAC1 receptors (52). Moreover, many endometrial carcinomas also have PAC1 receptors (52). Interestingly, medullary thyroid cancers are among the rare tumors that do not express VIP/PACAP receptors (186). Although VPAC1 mRNA was identified in many tumor cell lines, such as lung and breast cancers (Refs. 190 and 191 ; for review, see Ref.174), immunohistochemical data showing VIP/PACAP receptor expression in human tumors are not yet available.

C. Cholecystokinin (CCK) receptors
1. CCK/gastrin and their receptor subtypes.
The gastrointestinal peptides gastrin and CCK exist in different molecular forms (Table 1Go). Pro-gastrin and pro-CCK can be processed to peptides of variable length but, as biologically active peptides, they have the same five terminal amino acids at their carboxy terminus. They act as neurotransmitters in the brain, as regulators of various functions of the gastrointestinal tract, primarily at the level of the stomach, pancreas, and gallbladder (192). In addition, they can act as physiological growth factors in most parts of the gastrointestinal tract (193, 194) and also as growth factors in several neoplasms, such as in colonic, gastric, and brain cancers (195, 196, 197, 198). CCK and gastrin actions are mediated by several receptor subtypes, the best characterized being CCK1 (formerly CCK-A) and CCK2 (formerly CCK-B) receptors (199, 200) (Table 1Go). They can be distinguished pharmacologically by their low (CCK1) vs. high (CCK2) affinity for gastrin, or by their different affinity for nonpeptidic selective CCK antagonists. Recently, additional CCK receptors have been described, such as a CCK-C or a gastrin receptor in Swiss 3T3 fibroblasts (201, 202, 203). Extensive information on CCK and gastrin signaling, recently reviewed (204), has been obtained over the past several years. CCK1 and CCK2 receptors have been identified in several normal tissues (for review, see Ref.205). CCK2 receptors are present predominantly in the gut mucosa, in the endocrine pancreas, and in the brain (206, 207, 208); CCK1 receptors in the gallbladder, in gastric smooth muscles (208, 209), and in the peripheral nervous system, for instance in afferent vagal neurons (210) or in the myenteric plexus (182). As reported recently (211), human pancreatic acinar cells do not express a significant amount of CCK receptors in contrast to rat pancreatic acinar cells. This is a further example of the wide species variability of peptide receptor expression. CCK receptors, as most peptide receptors, can be rapidly internalized (23).

2. CCK1 and CCK2 receptors in cancer.
It has been established for a long time that small cell lung cancers often express CCK2 receptors, whereas non-small cell lung cancers do not (51, 212, 213). The findings are more equivocal for gastrointestinal cancers (214). Whereas earlier studies have reported the presence of CCK2 receptors in carcinomas of colon and stomach (215), more recent investigations have failed to find high-affinity CCK2 receptor proteins in most of these tumors (45, 216) although their mRNA is usually identified (217). The same may also be true for exocrine pancreatic carcinomas; although CCK1 and CCK2 receptor mRNA were identified in most tumors (218, 219), the receptor protein is difficult to detect in tumor cells themselves. The most frequently identified CCK receptor-expressing tissue elements in pancreatic cancer samples are nerves (CCK1) and islets (CCK2), but not tumor cells (219A ). A possible explanation for some of these discrepancies may be the existence of CCK2 receptor mutations in pancreatic, colorectal, and gastric cancers (220, 221, 222). A misspliced form of the CCK2 receptor that was detected in these tumors has constitutive activity and trophic effects (220, 221). Such mutated receptors may have altered binding characteristics. Recently, however, a high incidence of the regular CCK2 receptor protein was identified in medullary thyroid carcinomas (92%), whereas it was absent in differentiated thyroid cancers (223, 224, 225). CCK2 receptors were also found frequently in astrocytomas (65%) and in sex cord-stromal ovarian cancers (100%) (51) (Table 4Go); in some of the neuroendocrine gastroenteropancreatic tumors (in particular insulinomas) (156); in breast and endometrial adenocarcinomas; and in several soft tissue tumors, in particular in leiomyosarcomas (226). They were either not expressed or rarely expressed in meningiomas, neuroblastomas, schwannomas, glioblastomas, lymphomas, renal cell cancers, prostate carcinomas, hepatocellular carcinomas, and neuroendocrine tumors such as pituitary adenomas, pheochromocytomas, paragangliomas, or parathyroid adenomas (Table 4Go). CCK1 receptors were expressed in neuroendocrine lung and gastroenteropancreatic tumors, meningiomas, and some neuroblastomas (51, 156, 227). Immunohistochemical detection of CCK1 or CCK2 receptors in tumors has not yet been reported.

Gastrin mRNA measured by in situ hybridization was found to be present in some CCK2 receptor-positive small cell lung cancers, breast tumors, ovarian tumors, and stem cell tumors of various origins, possibly as indicators of an autocrine growth regulation of these tumors (51, 226). Conversely, gastrin and CCK mRNAs were absent in CCK2-expressing medullary thyroid cancers.

D. Bombesin/gastrin-releasing peptide (GRP) receptors
1. Bombesin/GRP and their receptor subtypes.
Bombesin and GRP, members of a family of brain-gut peptides, play an important role in cancer (228, 229, 230), in addition to their physiological function. Bombesin is a 14-amino-acid peptide present in amphibian tissues, whereas GRP, its human counterpart, consists of 27 amino acids. GRP and bombesin differ by only one of the 10 carboxy-terminal residues. This explains the similar biological activity of the two peptides. GRP acts primarily in the central and enteric nervous systems where it regulates several physiological processes including satiety, thermoregulation, circadian rhythm, smooth muscle contraction, immune function, as well as the release of other peptide hormones (229, 230). However, of all of the effects of GRP, the most studied is the one related to cancer. It was observed several years ago that cancer cell lines as well as primary human tumors can synthesize bombesin and GRP (231). Cuttitta et al. (228) showed that bombesin and GRP can stimulate small cell lung cancer growth and that this action is part of an autocrine feedback mechanism involving the expression of these peptides and that of their receptors in the tumor cells (232, 233). More recently, GRP and bombesin were deemed to play a role in other cancers as well. Stimulation of proliferation by bombesin was reported for lung, breast and pancreatic cancers (234, 235, 236). Moreover, GRP can promote cell proliferation in neuroblastoma cell lines (237) or in the androgen-independent human prostatic carcinoma cell line PC3; antagonists to the GRP receptor inhibit the growth of human prostatic carcinoma or of glioblastoma xenografts in nude mice (238, 239). Bombesin and GRP mediate their actions through membrane-bound, G protein-coupled receptors, which include at least four different subtypes, namely the neuromedin B receptor subtype (BB1), the GRP receptor subtype (BB2), the BB3 and BB4 subtypes (240, 241, 242, 243). With the exception of the GRP receptor (182, 244), these subtypes have been poorly characterized in regard to their distribution and function in human tissues.

2. Bombesin/GRP receptors in human tumors.
Although GRP receptors have been readily detected in various types of tumor cell lines (245, 246, 247), it has been more difficult to identify them in primary human cancers. GRP receptor mRNA could well be measured in various human neoplasms, including cancers of the gastrointestinal tract, lung, prostate, and breast (248, 249, 250). Recently, GRP receptors have also been detected in neuroblastomas by immunohistochemistry (237). The GRP receptor proteins have been difficult to detect with binding methods in gastrointestinal cancers (251, 252). Results have been controversial in exocrine pancreatic carcinomas: whereas one study found GRP-receptor expression in these cancers (253), more recent investigations identified these receptors extremely rarely (254, 255); however, peritumoral vessels surrounding exocrine pancreatic carcinomas clearly express GRP receptors (254). GRP receptor proteins have been more easily identified in renal cell, breast, and prostate carcinomas (248, 251, 256, 257, 258). Interestingly, some tissues of origin of these cancers express the GRP receptor (e.g., breast; Ref.256), whereas others do not (e.g., prostate; Ref.257).

Two possible reasons have been put forward to explain the difficulty of detecting GRP receptor proteins in some tumors: some authors have suggested that the potent neutral endopeptidase EC 3.4.24.11 (259) rapidly degrades the bombesin ligands (251); others have proposed that aberrant and possibly mutated GRP receptors (249, 260) with altered pharmacological characteristics (261) account for this difficulty.

Because of the potential clinical impact, it is worth emphasizing the strong GRP receptor expression in two tumoral conditions. GRP receptors were detected, often in high density: 1) in 30 of 30 invasive prostatic carcinomas and in 26 of 26 cases of prostatic intraepithelial proliferative lesions, mostly prostatic intraepithelial neoplasias (257). Bone metastases of androgen-independent prostate cancers were also GRP receptor-positive in four of seven cases. Conversely, GRP receptors, absent in normal prostate, were identified in only a few hyperplastic prostates; they were localized in very low density in glandular tissue and, focally, in some stromal tissue (257). The massive GRP receptor expression in prostate tissues that are in the process of malignant transformation (e.g., in prostatic intraepithelial neoplasias) or that are completely neoplastically transformed suggests that GRP receptors may be markers for early molecular events in prostate carcinogenesis and useful in differentiating prostate hyperplasia from prostate neoplasia. 2) GRP receptors were also detected in neoplastic epithelial mammary cells in two thirds of invasive ductal carcinomas and ductal carcinomas in situ (256). The lymph node metastases from those primary carcinomas expressing GRP receptors were all positive, whereas surrounding lymphoreticular tissue was GRP receptor-negative. Although these receptors were also present in ducts and lobules from nonneoplastic breast tissue samples, the strong GRP receptor expression in breast carcinomas suggests that these tumors may be a consequential target for GRP and bombesin analogs (256).

Recently, a very potent ligand, the [D-Tyr6, ß-Ala11, Phe13, Nle14]bombesin(6, 7, 8, 9, 10, 11, 12, 13, 14), shown to be bound by all four bombesin receptors, has been developed by the Jensen group (262, 263). This compound, iodinated at the D-Tyr6 residue, yields a useful radioligand able to distinguish the various bombesin receptor subtypes on the basis of the rank order of their affinity for GRP, neuromedin B, [D-Tyr6, ß-Ala11, Phe13, Nle14]bombesin (6, 7, 8, 9, 10, 11, 12, 13, 14), or bombesin. Using this approach, we could specifically detect BB3 receptor subtype expression in human pancreatic islets (254). More recently, we identified neuroendocrine tumors with a differentiated profile of receptor subtype expression: gastrinomas displayed preferentially GRP receptors, and ileal carcinoids expressed often neuromedin B receptors, whereas bronchial carcinoids and small cell lung carcinomas frequently had BB3 receptors (156, 264).

E. Neurotensin receptors
1. Neurotensin and its receptor subtypes.
Neurotensin is a tridecapeptide localized both in the central nervous system and in peripheral tissues, mainly in the gastrointestinal tract (265, 266, 267). In the central nervous system, neurotensin plays the role of neurotransmitter or neuromodulator of dopamine transmission and of anterior pituitary hormone secretion (267). It also shows potent hypothermic and analgesic effects in the brain. In the periphery, neurotensin acts as a local hormone exerting a paracrine and endocrine modulation of the digestive tract (267). It may also play an important role in gut mucosal immune responses (268, 269). Finally, it can stimulate growth in a variety of normal cells (270). The pharmacological effects of neurotensin result from the specific interaction of the peptide with cell-surface receptors. However, the pharmacology and mode of action of neurotensin receptors are not completely clear. The signaling pathway of one of the neurotensin receptor subtypes, the high-affinity NTR1 receptor, is well documented, including Ca2+ release after inositol 1,4,5-triphosphate stimulation (271), activation of MAPKs (272) via protein kinase C, leading to its role in cell proliferation. All of the effects mediated by the NTR1 are blocked by its selective nonpeptide antagonist SR 48692 (273). By contrast, the signaling pathway that governs the interaction of neurotensin with another neurotensin receptor subtype, the levocabastine-sensitive NTR2 receptor (274), is a matter of controversy. The complexity of neurotensin signaling has been recently emphasized by the molecular identification of a third membrane protein, non-G protein-coupled, capable of binding the peptide with a high affinity, the NTR3 receptor (275). This protein is identical to gp95 sortilin, a sorting protein originally identified by its ability to interact with a receptor-associated protein (276). Neurotensin receptors have been shown to be internalized after interaction with the peptide (24, 277). For instance, after interaction with neurotensin, 60–70% of the NTR1 receptor present in COS cells internalizes according to a temperature-dependent process (271, 278). Neurotensin is rapidly degraded in blood plasma by endogenous peptidases and proteases. Several proteolytic enzymes including neutral endopeptidase EC 3.4.24.11, angiotensin-converting enzyme, and metalloendopeptidases EC 3.4.24.15 and EC 3.4.24.16 have been reported to cleave intact neurotensin.

2. Neurotensin receptors in cancer.
Several lines of evidence suggest that neurotensin plays a role in cancer. It is known that neurotensin receptors are expressed in various tumor cell lines including small cell lung cancer, neuroblastoma, pancreatic, or colonic cancer (279, 280). Clinically more relevant, they can be overexpressed in primary human tumors, e.g., in most meningiomas and Ewing’s sarcomas (281, 282), more than three fourths of ductal pancreatic carcinomas (255, 283), and, in a somewhat lower incidence, in astrocytomas, medulloblastomas, medullary thyroid cancers, and small cell lung cancers (282). These neoplasms display NTR1 receptor proteins, characterized by their low affinity for levocabastine, as well as NTR1 mRNA (284, 285). NTR1 was rarely found in non-small cell lung cancers; in carcinomas of the breast, colon, rectum, prostate, ovary, renal, or hepatic cells; in neuroendocrine gut tumors; pituitary adenomas; schwannomas; neuroblastomas; and lymphomas (45, 282) (Table 4Go). An additional argument in favor of a role of neurotensin in cancer is that this peptide can stimulate the proliferation in vitro of tumor cell lines of various origins, including those originating in the pancreas, prostate, brain, and lung (30, 197, 286, 287). Conversely, the NTR1 receptor antagonist SR 48692 (273) inhibits tumor proliferation (30, 286, 287, 288, 289). Also of further interest is NTR3, which has recently been shown to be involved in mediating neurotensin growth stimulation in cancer cell lines (290). The complex neurotensin-mediated signal transduction mechanisms are presently under investigation in cancer models (291, 292).

Interestingly, neurotensin itself appears to be expressed by numerous tumors or tumor cell lines. For instance, in some receptor-positive Ewing’s sarcomas, neurotensin mRNA was detected by in situ hybridization techniques (282). The presence of neurotensin and of the neurotensin receptor in human neoplasia may therefore be an integrative part of an autocrine feedback mechanism of tumor growth stimulation (282), as shown previously for the GRP system.

F. Other peptide receptors
1. Substance P.
Substance P is a neuropeptide involved in a variety of functions of the central and peripheral nervous systems, including pain perception and vasodilatation (293). One of the substance P receptor subtypes, the NK1 receptor (Table 1Go), is quite frequently expressed in glial tumors, in particular in poorly differentiated glioblastomas, but can also be detected in medullary thyroid carcinomas, small cell lung cancers, pancreatic as well as breast cancers (294, 295); it is rarely found in gastrointestinal tumors or lymphomas. Interestingly, all tumor types, regardless of their histology, show high levels of NK1 receptor expression in tumoral and peritumoral vessels. These vascular receptors may serve as the molecular basis for a substance P-mediated vasodilatation (294, 296). Substance P is able to stimulate the proliferation of malignant tumor cells (295). Accordingly, NK1 receptor antagon