Internalized Somatostatin Receptor Subtype 2 in Neuroendocrine Tumors of Octreotide-Treated Patients

Abstract

Context: Somatostatin receptor subtype 2 (sst₂) is widely expressed in neuroendocrine tumors and can be visualized immunohistochemically at the cell membrane for diagnostic purposes. Recently, it has been demonstrated in animal sst₂ tumor models in vivo that somatostatin analog treatment was able to induce a complete internalization of the tumor sst₂.

Patients and Methods: In the present study, we evaluated whether sst₂ expressed in neuroendocrine tumors of patients treated with octreotide are also internalized. Tumor samples were assessed in patients that were treated with various octreotide modalities before and during surgery and compared with tumor samples from untreated patients. Sst₂ immunohistochemistry was performed in all samples with three different sst₂ antibodies (R2-88, UMB-1, and SS-800). Sst₂ receptor expression was confirmed by immunoblotting and in vitro receptor autoradiography.

Results: Patients receiving a high dose of octreotide showed predominantly internalized sst₂, and patients with a low dose of octreotide had a variable ratio of internalized vs. membranous sst₂, whereas untreated patients had exclusively membranous sst₂. The internalized sst₂ receptor corresponded to a single sst₂ band in immunoblots and to sst₂ receptors in in vitro receptor autoradiography. Although generally found in endosome-like structures, internalized sst₂ receptors were also identified to a small extent in lysosomes, as seen in colocalization experiments.

Conclusion: It is the first evidence showing that sst₂ receptors can be internalized in sst₂-expressing neuroendocrine tumors in patients under octreotide therapy, providing clues about sst₂ receptor biology and trafficking dynamics in patients.

Somatostatin sst₂ receptors can be internalized in sst₂-expressing neuroendocrine tumors in patients under octreotide therapy, providing clues about sst₂ receptor biology and trafficking dynamics in patients.

One of the general physiological feature of G protein-coupled receptors is that agonist binding to the receptor triggers an internalization of the receptor-ligand complex into the cells (1,2,3,4). This occurs physiologically after binding of endogenous ligand to the receptor, as shown in the elegant study by Mantyh et al. (5,6) on substance P receptor internalization after nerve stimulation. It also occurs after binding of exogenous ligands to the receptor, as illustrated by the field of peptide receptor tumor targeting, where receptor-mediated internalization of radiolabeled peptides is used for diagnostic and therapeutic applications (7,8,9). The best examples in this regard are the somatostatin receptors: the high degree of internalization of somatostatin radioligands into somatostatin receptor subtype 2 (sst₂)-expressing tumor cells is a powerful mechanism of radioactivity accumulation within the tumor that may permit the successful imaging of tumors in patients as well as targeted tumor radiotherapy (8,10,11).

Although most of the internalization studies have been performed in vitro (12,13,14,15), it was reported recently by Waser et al. (16) that internalization of the sst₂ receptor could also be observed in vivo in animal tumor models after iv application of the somatostatin agonist [Tyr³] octreotate. It was found that the agonist-induced sst₂ internalization was extremely rapid and powerful and that almost all sst₂ receptors moved from the cell membrane to endosome-like cellular structures within the cytoplasm, giving a characteristic immunohistochemical pattern of receptor distribution (16).

Because the in vivo sst₂ internalization process after agonist application is so potent and prominent, we were interested to know whether internalized sst₂ could also be detected in resected tumor tissue from patients treated with the somatostatin agonist octreotide immediately before or during surgery. Using sst₂ immunohistochemistry, a method shown previously to identify sst₂ receptors on formalin-fixed, paraffin-embedded tumors (17,18,19,20,21), we have analyzed in the present study the tumor tissue samples from patients treated with octreotide [including octreotide long-acting repeatable (LAR), octreotide infusion during surgery, or sc octreotide application at beginning of surgery] and compared with tumor samples from patients that had not been treated with octreotide. Additional confirmatory proofs for sst₂ expression in these samples were obtained with in vitro somatostatin receptor autoradiography and immunoblotting.

Materials and Methods

Cell line

The HEK293 cell line expressing the human T7-epitope-tagged sst₂ receptor (HEK-sst₂) was cultured at 37 C and 5% CO₂ in DMEM with GlutaMax I containing 10% (vol/vol) fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, and 500 μg/ml G418. All culture reagents were from Life Technologies, Inc. (Grand Island, NY).

Tissues

Formalin-fixed and fresh-frozen tissue samples from surgically resected neuroendocrine tumors expressing sst₂ receptors were used. The samples were divided into three groups depending on the type of octreotide treatment received during surgical tumor resection: 1) tumors from patients that had not been in contact with octreotide before or during surgery, 2) tumors from patients that had received 200 μg octreotide sc at the start of surgery, and 3) tumors from patients that had received an iv octreotide infusion (200 μg/h) during the time of surgical resection plus 20 mg octreotide LAR less than 3 wk before surgery. For each patient, a specific clinical rationale was the basis for the particular perioperative octreotide treatment, namely to avoid preoperative complications in functional neuroendocrine tumors (22) or to improve the postoperative outcome of pancreatic interventions as suggested by a previous study (23). Limitations due to patient-based and surgery-dependent type of investigation included variable route of application of octreotide (iv injection, sc injection, and LAR) and variable intervals between octreotide injection and tumor removal. Written informed consent was available for all patients. The present study conformed to the ethical guidelines of the Institute of Pathology of the University and the University Hospital of Berne and was reviewed by the Institutional Review Board.

Sst₂ immunohistochemistry

Sst₂ receptors were detected using immunohistochemistry, as described before (16,17,18,24,25). All samples were tested with the polyclonal sst₂-specific antibody R2-88 (generously provided by Dr. A. Schonbrunn, Houston, TX), with the commercially available polyclonal sst₂-specific antibody SS-800 (Gramsch Laboratories, Schwabhausen, Germany) and with the rabbit monoclonal antibody UMB-1 (SS-8000RM; Biotrend GmbH, Cologne, Germany) (17,21). Four-micrometer-thick formalin-fixed, paraffin-embedded tissue sections were used. As reported previously (18), the best antigen retrieval method for R2-88 and UMB-1 immunohistochemistry was boiling in the microwave in presence of a 5% urea buffer (pH 9.5). For SS-800 immunohistochemistry, boiling in the microwave in the presence of a 10 mm citrate buffer (pH 6.0) was preferable (17,21). Both R2-88 and SS-800 were applied in a 1:1000 dilution, whereas a 1:100 dilution was necessary for UMB-1. The secondary antibody was a biotinylated goat antirabbit Ig (1:200 dilution). Antibody binding was visualized using the Vectastain Elite ABC kit (Vector Laboratories, Burlingame, CA). Staining was carried out with diaminobenzidine and counterstaining with hemalum. For specificity control, the primary antibodies were preadsorbed with 100 nm of the corresponding antigen peptide. In all experiments, a well-characterized gastro-enteropancreatic neuroendocrine tumor strongly expressing sst₂ as determined by receptor autoradiography was included as positive control. As controls for membranous or internalized sst₂, AR42J tumor tissue and pancreas from rats injected either with saline or with [Tyr³] octreotate, respectively (16), were analyzed in parallel.

Receptor autoradiography

In all cases with available fresh-frozen tumor tissues, somatostatin receptor autoradiography was carried out as described before (26). Briefly, 20-μm-thick frozen tissue sections were cut from each of the collected tumor samples and were incubated for 2 h at room temperature with the sst₂-preferring somatostatin radioligand ¹²⁵I-labeled [Tyr³] octreotide (26). Nonspecific binding was determined in serial tissue sections incubated with the radioligand in the presence of 100 nm octreotide. The slides were subsequently exposed to Kodak Biomax MR films for 7 d at 4 C. Receptor quantification was performed as previously described (26).

Immunoblotting

Approximately 50–100 mg fresh-frozen neuroendocrine tumors was cut into small pieces, partially thawed, and washed in ice-cold PBS containing protease inhibitors (1 mm phenylmethylsulfonyl fluoride, 10 μg/ml soybean trypsin inhibitor, 10 μg/ml leupeptin, 5 μg/ml aprotinin, and 50 μg/ml bacitracin), incubated for 30 min at 4C in ice-cold homogenization buffer [20 mm HEPES, 150 mm NaCl, 5 mm EDTA, 3 mm EGTA, and 4 mg/ml dodecyl β-d-maltoside (pH 7.6)] containing protease inhibitors, and then homogenized and lysed in a precooled Dounce homogenizer. The lysate was subsequently clarified by centrifugation at 16,000 × g and 4 C for 30 min and then subjected to partial purification by wheat germ agglutinin-agarose treatment. The adsorbed glycoproteins were eluted at 60 C for 15 min in SDS-PAGE sample buffer containing 10% 2-mercaptoethanol and then resolved on a 10% SDS-polyacrylamide gel. Resolved proteins were transferred to polyvinylidene difluoride membrane. The membrane was blocked for 2 h with 5% nonfat dry milk in Tris buffered saline-Tween [TBS-T: 20 mm Tris-HCl, 150 mm NaCl, 0.1% Tween 20 (pH 7.4)], and then incubated over night at 4 C with the sst₂-specific antibody R2-88 (1:10,000) in TBS-T. Immunoreactive proteins were detected with a donkey antirabbit-F(ab′)₂ conjugated with horseradish peroxidase (1:10,000) in TBS-T, and the ECL chemiluminescent antibody detection system (GE Healthcare, Little Chalfont, UK). As positive control, lysates of HEK-sst₂ cells were used and treated as described above for the lysates of the neuroendocrine tumors.

Confocal double-labeling immunofluorescence: histochemistry

Selected sst₂-expressing neuroendocrine tumor samples were used for colocalization studies using confocal immunohistochemistry. The samples were analyzed for colocalization with the sst₂-specific antibody R2-88 and with a monoclonal antibody raised against the lysosomal marker protein cathepsin D (ab6319; Abcam Ltd., Cambridge, UK). Four-micrometer-thick formalin-fixed, paraffin-embedded tissue sections were used and treated as described above for sst₂ immunohistochemistry. The best antigen retrieval method for R2-88 and anti-cathepsin D antibody immunohistochemistry was boiling in the microwave in the presence of a 5% urea buffer (pH 9.5). Both antibodies, R2-88 and the anti-cathepsin D antibody, were applied together in a dilution of 1:1000 overnight at room temperature. The secondary antibody for R2-88 was a goat antirabbit IgG Alexa 488 conjugate (Molecular Probes, Eugene, OR), whereas the secondary antibody for the anti-cathepsin D antibody was a goat antimouse IgG Alexa 594 conjugate (Molecular Probes). Both secondary antibodies were applied together in a dilution of 1:300 for 45 min at room temperature and in the dark. Slides were subsequently mounted with a polyvinyl-alcohol- and glycerine-based mountant containing 1,4-diazabicyclo(2,2,2)octane (polydabco). Confocal imaging was performed on an Olympus FluoView FV1000 scanning unit with the IX81 inverted microscope (Olympus, Tokyo, Japan). All confocal images processed for analysis were collected using a ×60 oil immersion lens.

Confocal double-labeling immunofluorescence: cytochemistry

HEK-sst₂ cells were grown on poly-d-lysine-coated (20 μg/ml) (Sigma-Aldrich, St. Louis, MO) 10-mm glass coverslips in 24-well plates for 24 h at 37 C and with 5% CO₂. Cells were then treated with 1 μm [Tyr³] octreotide (Novartis, Basel, Switzerland) at 37 C and with 5% CO₂ in growth medium for 30 min and then processed for immunofluorescence microscopy as described previously (12,13) using as primary antibodies the sst₂-specific antibody R2-88 and the monoclonal antibody raised against the lysosomal marker protein cathepsin D (ab6319; Abcam Ltd., Cambridge UK). Both antibodies, R2-88 (1:1000 dilution) and the anti-cathepsin D antibody (1:500 dilution), were applied together for 1 h at room temperature. The secondary antibody for R2-88 was a goat antirabbit IgG Alexa 488 conjugate (Molecular Probes), whereas the secondary antibody for the anti-cathepsin D antibody was a goat antimouse IgG Alexa 594 conjugate (Molecular Probes). Both secondary antibodies were applied together in a dilution of 1:600 for 45 min at room temperature and in the dark. Slides were subsequently mounted with polydabco. Confocal imaging was performed on an Olympus FluoView FV1000 scanning unit with the IX81 inverted microscope (Olympus, Tokyo, Japan). All confocal images processed for analysis were collected using a ×60 oil immersion lens.

Results

The immunohistochemical data are summarized in Table 1 and in Figs. 1, 2, and 4. Tumor samples of patients that had not received an octreotide therapy exhibited exclusively a membranous sst₂ staining. In contrast, the tumor samples from octreotide-treated patients showed internalized sst₂ receptor in variable amounts (Table 1 and Figs. 1 and 2). The two cases (patients 1 and 2) with 200 μg/h octreotide infusion and 20 mg octreotide LAR therapy, reflecting a high octreotide dose (22), showed predominantly internalized sst₂ receptor with few focal areas with membranous sst₂. The cases with sc octreotide application (200 μg single injection) at the beginning of the surgery showed internalized sst₂ as well, but the ratio between internalized and membranous sst₂ varied considerably between individuals. Some cases had predominantly internalized sst₂ receptors, others a mixture of internalized and membranous receptors, and few had preferably membranous receptors (Table 1). Moreover, as reported previously for octreotide-naive patients (17), the present study repeatedly identified heterogeneous receptor-expressing tumor areas, i.e. containing sst₂-positive tumor cells adjacent to sst₂-negative ones. This is illustrated in Fig. 1 (D–F) in samples of octreotide-treated patients, showing adjacent tumor cells with and without internalized sst₂. To provide strong immunohistochemical proof, we used three different sst₂-specific antibodies, recognizing C-terminal sst₂ epitopes on successive sections of each of the tumor samples. Comparative qualitative results were obtained for each of the three antibodies. Figure 1 shows examples of neuroendocrine tumor samples revealing a comparable immunostaining with R2-88, UMB-1, or SS-800 in adjacent sections, with sst₂ receptors internalized in two octreotide-treated cases and a membranous sst₂ staining in one control sample.

Table 1.

Immunohistochemical data

Patients	Tumor type	OCT long-term therapy	OCT pre-/peri-OP	Sst2 IHCa	ARG
Octreotide-treated
1	NET (pancreatic)	LAR 20 mg 3 wk pre-OP	200 μg/h infusion peri-OP	Mostly internal. sst2	+++
2	NET (bronchial; heart-met)	LAR 20 mg 3 wk pre-OP	200 μg/h infusion peri-OP	Mostly internal. sst2	NT
3a	NET (pancreatic)	No	200 μg sc	Mostly internal. sst2	+++
3b	NET (pancreatic)	No	200 μg sc	Mostly internal. sst2	+++
4a	NET (pancreatic)	No	200 μg sc	Internal. = membr. sst2	NT
4b	NET (pancreatic)	No	200 μg sc	Internal. = membr. sst2	NT
5a	NET (pancreatic; primary)	No	200 μg sc	Internal. = membr. sst2	+++
5b	NET (pancreatic; liver-met)	No	200 μg sc	Internal. = membr. sst2	+++
6	NET (pancreatic)	No	200 μg sc	Membr. > internal. sst2	+++
7	NET (pancreatic)	No	200 μg sc	Mostly membr. sst2	+++
8	NET (pancreatic)	No	200 μg sc	Mostly membr. sst2	+++
9	NET (pancreatic)	No	200 μg sc	Mostly membr. sst2	+++
Untreated
10	NET (bile duct)	No	No	Membr. sst2	++
11a	NET (pancreatic)	No	No	Membr. sst2	+++
11b	NET (pancreatic)	No	No	Membr. sst2	+++
12	NET (bronchial)	No	No	Membr. sst2	NT
13	NET (mesenteric)	No	No	Membr. sst2	+++
14	NET (pancreatic; liver-met)	No	No	Membr. sst2	+++
15	NET (pancreatic)	No	No	Membr. sst2	+++

ARG, Somatostatin receptor autoradiography; IHC, immunohistochemistry; internal., internalized; membr., membranous; NET, neuroendocrine tumor; OCT, octreotide; OP, operation; met, metastasis; NT, not tested. +++ indicates high density of somatostatin receptors (>8000 dpm/mg tissue); ++ represents a density of 4567 dpm/mg tissue. 

^aAll three sst₂ antibodies (R2-88, UMB-1, and SS-800) were tested in each case with comparable results. 

Figure 1.

Sst₂ immunohistochemistry with three different sst₂-specific antibodies: R2-88 (A, D, and G), UMB-1 (B, E, and H), and SS-800 (C, F, and I). Bars, 0.01 mm. A–C, Tumor tissue from patient 11b not treated with octreotide (control) showing membranous sst₂ staining with all three antibodies. D–F, Tumor tissue from patient 2 treated with 20 mg octreotide LAR plus 200 μg/h octreotide infusion perioperatively, showing intracellular sst₂ staining with all three antibodies. G–I, Tumor tissue from patient 3a treated with 200 μg octreotide sc at start of surgery, showing intracellular sst₂ staining with all three antibodies.

Figure 2.

Sst₂ immunohistochemistry with R2-88 (columns A and B) in neuroendocrine tumor samples from four patients treated with octreotide is compared with sst₂ expression measured in the same tissue samples by in vivo receptor autoradiography (columns C and D). Column A shows the great individual variability of sst₂ internalization among the four cases, ranging from predominantly internalized sst₂ (top) to predominantly membranous sst₂ (bottom). Bars, 0.01 mm. Column B shows that the sst₂ immunostaining is specific in all cases because preadsorption with excess antigen peptide abolishes all staining. Column C shows autoradiograms showing total binding of ¹²⁵I-labeled [Tyr³] octreotide in the four different tumors. All show a very high density of somatostatin receptors. Bars, 1 mm. Column D shows autoradiograms showing nonspecific binding of ¹²⁵I-labeled [Tyr³] octreotide (in the presence of 10⁻⁶ m octreotide). The four rows from top to bottom correspond to the samples of patients 3a, 1, 5b, and 8, respectively.

Figure 4.

Sst₂ immunohistochemistry showing internalized sst₂ receptors at high magnification in tumors from three octreotide-treated patients (A–C) as compared with a tumor (D) of an untreated patient showing membranous staining. For comparison, E shows internalized sst₂ in an AR42J tumor of a rat treated with 0.21 mg [Tyr³] octreotate given iv 1 h before being killed. Bars, 0.01 mm.

Figure 2 illustrates the inter-individual variability of sst₂ internalization and documents the specificity of the immunohistochemical observation. Proof for the specificity of the R2-88 immunostaining in all those cases is given by the control samples showing suppression of the internalized as well as of the membranous immunostaining after adsorption with a high dose of the antigen peptide (Fig. 2). This proof of specificity was also obtained with the two other sst₂ antibodies (UMB-1 and SS-800) after antigen peptide treatment. Furthermore, Fig. 2 shows that the strong sst₂ receptor expression in the four positively immunostained tumor samples can be confirmed with a completely different method of receptor detection, namely in vitro receptor autoradiographic identification using the sst₂-preferring radioligand ¹²⁵I-labeled [Tyr³] octreotide. Confirmation of sst₂ expression by receptor autoradiography was obtained for all cases where fresh tissue was available (Table 1).

Figure 3 demonstrates in three cases with predominantly internalized sst₂ receptors (octreotide-treated patients) that the observed positive R2-88 immunohistochemical staining corresponds to a specific band on an immunoblot using the R2-88 antiserum. A corresponding sst₂ band is also seen in a case with membranous sst₂ immunohistochemical staining (untreated patient) and in the sst₂-expressing HEK-sst₂ cell line, used as controls. As specificity control, all bands are shown to be abolished in the presence of the antigen peptide (Fig. 3).

Figure 3.

Confirmation by immunoblotting (bottom row) of the sst₂ expression detected with sst₂ immunohistochemistry (top row) in tumor samples of octreotide-treated patients (patients 1, 5b, and 3a) compared with controls consisting of sst₂-expressing HEK293 cells (HEK-sst₂) and a tumor sample from a patient not treated with octreotide (patient 15). In the immunoblot, a broad band centered around 70 kDa is stained by the R2-88 antiserum in all tested samples. This staining is completely abolished by preadsorption of R2-88 with the antigen peptide (a.p.). The variation in apparent molecular mass of the sst₂ receptor in the analyzed samples reflects the differences of glycosylation of the receptor. The positive sst₂ staining detected with the R2-88 antiserum in the immunohistochemistry samples of patients 1, 5b, and 3a as well as the membranous staining seen in the sample of patient 15 together with the presence of a sst₂ receptor band in all samples detected by R2-88 immunoblotting strongly suggests that the granular cytoplasmic structures seen by immunohistochemistry indeed correspond to the internalized sst₂ receptor. Pat., Patient.

Figure 4 shows at high magnification the localization pattern of the internalized sst₂ receptor compared with the membranous sst₂ location in an patient not treated with octreotide. The cytoplasmic pattern is compatible with an endosomal localization of the sst₂ receptor, as observed in AR42J tumors of [Tyr³] octreotate-treated rats (Fig. 4E). Figure 5 illustrates by confocal double-labeling immunofluorescence microscopy that a small amount of the internalized sst₂ receptor is routed from the endosomal system to lysosomes for degradation. This is confirmed by the colocalization of a small population of sst₂ receptors, detected by R2-88 antiserum, with the lysosomal marker protein cathepsin D in samples from octreotide-treated patients as well as in the agonist-stimulated control cell line HEK-sst₂.

Figure 5.

Comparison of localization of internalized sst₂ receptor with the lysosomal marker protein cathepsin D in tumor samples of octreotide-treated patients (top row, patient 5b; middle row, patient 3b) compared with [Tyr³] octreotide-treated, sst₂-expressing HEK293 cells (bottom row, HEK-sst₂) by confocal double-labeling immunofluorescence imaging. The left column (green) shows the internalized sst₂ receptor visualized by the R2-88 antiserum, and the middle column (red) shows the lysosomal staining of cathepsin D. The merged images are shown in the right column. In both tissue samples and in the HEK-sst₂ cells, the sst₂ receptor shows only a minimal overlap with cathepsin D in the merged images (yellow structures), suggesting that upon agonist-stimulated sst₂ receptor internalization, only few sst₂ receptors are localized in lysosomal structures.

Discussion

This is the first study indicating that octreotide treatment is associated with agonist-induced internalization of sst₂ receptors in human sst₂-expressing neuroendocrine tumors. The degree of internalization, i.e. the ratio of internalized sst₂ to membranous sst₂, varies very much from one patient to the other. A particularly powerful, high-dose octreotide treatment combining octreotide LAR and octreotide infusion (22) during surgical resection triggers a strong and almost complete sst₂ internalization. Conversely, with 200 μg sc octreotide given at the start of surgery, a wide range of sst₂ internalization pattern is observed, ranging from almost complete receptor internalization in some cases to only focal or even absent internalization next to a predominantly membranous sst₂ immunostaining in other individuals. Also, a heterogeneous distribution of tumor cells expressing the internalized receptors may be observed, with cells expressing the sst₂ receptors located next to cells without receptors, as reported previously for cells with membrane-bound sst₂ (17). The above mentioned diversity in internalization patterns may be due to the fact that the 200 μg sc octreotide is a submaximal dose (22) that may not be sufficient to trigger a complete and long-lasting sst₂ internalization in all tumors. The octreotide concentration at the target will indeed depend on numerous systemic and local factors, including the performance of the cardiovascular system as well as local tumor perfusion, tumor sst₂ receptor density, and finally, the time elapsing between sc application and tumor resection.

The arguments that the observed cytoplasmic sst₂ immunostaining represents true sst₂ receptors are 4-fold. First, we observe a comparable immunohistochemical pattern with three distinct sst₂-specific antibodies, namely R2-88, SS-800, and UMB-1, that identify C-terminal epitopes. Second, the sst₂ immunostaining is always abolished in the presence of an excess of the respective peptide antigens. Third, the samples revealing intracellular immunostaining are found to express a high density of sst₂ receptors, as measured by in vitro receptor autoradiography with the sst₂-preferring radioligand ¹²⁵I-labeled [Tyr³] octreotide. Interestingly, all tested samples, including those with predominantly internalized sst₂, were found to be receptor positive in autoradiography, as opposed to a previously published single case of somatostatinoma expressing internalized sst₂ unidentified by autoradiography (25). Fourth, cases with predominantly internalized sst₂ detected by R2-88 immunostaining were submitted to R2-88 immunoblotting and found to exhibit a specific band corresponding to the predicted molecular mass of sst₂, strongly indicating that the cytoplasmic granular staining seen by immunohistochemistry indeed represents specific sst₂ receptors.

At a high magnification, the internalized sst₂ show a cytoplasmic distribution pattern compatible with localization in endosomal structures (12,13,27,28) and comparable with that of internalized sst₂ in AR42J tumors from [Tyr³] octreotate-treated rats, as reported previously (16). Furthermore, it appears that the internalized sst₂ are not exclusively localized in endosomal structures but that they can be found in smaller amounts also in lysosomes, as shown in the double-labeling immunofluorescence experiments detecting sst₂ and the lysosomal marker protein cathepsin D. This suggests that although the internalized sst₂ receptors are to a large amount recycled back to the cell surface, as shown previously by the reversibility of the sst₂ internalization process in AR42J tumors (16), some of the receptors will nevertheless be routed to the lysosomes and degraded.

Due to the availability of suitable sst₂ antibodies, there has been an increased use of sst₂ immunohistochemistry for the diagnostic evaluation of sst₂ expression in neuroendocrine tumors (17,18,19,21). Up to now, a membranous sst₂ localization was considered an accepted criterion for specific sst₂ expression and was usually required as a reliable diagnosis of sst₂ receptor expression in a tumor (17,18,19,20,21). The present data strongly suggest that the specific pattern of internalized sst₂ receptors should also be considered as an additional reliable criterion for specific sst₂ expression when it is confirmed that the patient has been on octreotide therapy immediately before or during surgery.

Sst₂ receptor desensitization and escape from octreotide treatment, in particular in gastro-enteropancreatic neuroendocrine tumors (29,30) has been an important clinical observation that has been repeatedly debated in the past (31,32,33,34,35). In addition, preoperative management of secreting neuroendocrine tumors is still a challenge. The present data are limited in resolving these problems, because octreotide treatment in the present study cannot be considered a long-term therapy, and the sst₂ data obtained on resected tumors represent only a single time point. This type of study, however, may open the possibility to investigate long-term effects of octreotide treatment on the cellular localization of sst₂ receptors in tumors in selected patients with and without escape from treatment.