Successful receptor-mediated radiation therapy of xenografted human midgut carcinoid tumour

Somatostatin receptor (sstr)-mediated radiation therapy is a new therapeutic modality for neuroendocrine (NE) tumours. High expression of sstr in NE tumours leads to tumour-specific uptake of radiolabelled somatostatin analogues and high absorbed doses. In this study, we present the first optimised radiation therapy via sstr using [177Lu-DOTA0-Tyr3]-octreotate given to nude mice xenografted with the human midgut carcinoid GOT1. The tumours in 22 out of 23 animals given therapeutic amounts showed dose-dependent, rapid complete remission. The diagnostic amount (0.5 MBq [177Lu-DOTA0-Tyr3]-octreotate) did not influence tumour growth and was rapidly excreted. In contrast, the therapeutic amount (30 MBq [177Lu-DOTA0-Tyr3]-octreotate) induced rapid tumour regression and entrapment of 177Lu so that the activity concentration of 177Lu remained high, 7 and 13 days after injection. The entrapment phenomenon increased the absorbed dose to tumours from 1.6 to 4.0 Gy MBq−1 and the tumours in animals treated with 30 MBq received 120 Gy. Therapeutic amounts of [177Lu-DOTA0-Tyr3]-octreotate rapidly induced apoptosis and gradual development of fibrosis in grafted tumours. In conclusion, human midgut carcinoid xenografts can be cured by receptor-mediated radiation therapy by optimising the uptake of radioligand and taking advantage of the favourable change in biokinetics induced by entrapment of radionuclide in the tumours.

In coupled with the somatostatin analogue octreotide via the chelator DTPA (Krenning et al, 1993;Ahlman et al, 1994;Cimitan et al, 2003). Pharmacological treatment of the carcinoid syndrome with long-acting somatostatin analogues reduces hormone release and related symptoms in the majority of patients. However, there is no evidence that somatostatin analogues exert antiproliferative actions on these tumours, although individual cases with induction of apoptosis have been reported (Eriksson and Ö berg, 1999;Bousquet et al, 2001). The only curative treatment for carcinoid tumours is surgical, but due to late diagnosis, radical surgery is seldom possible. Debulking procedures offer palliation of hormonal symptoms and may prolong survival (McEntee et al, 1990;Ahlman et al, 2000). Palliative treatment may include chemotherapy and biotherapy with interferon, both with limited benefits and adverse effects (Ö berg, 1993;Moertel et al, 1994;Kölby et al, 2003). Following interventional and medical treatment, the 5-year survival is as high as 69% for patients with liver metastases Hellman et al, 2002).
A new treatment strategy is directed radiation therapy with radiolabelled somatostatin analogues. The first therapies were attempted with [ 111 In-DTPA 0 ]-octreotide and resulted in reduced tumour markers such as chromogranin A (CgA, the general NE marker) and 5-hydroxy indole acetic acid (5-HIAA, the serotonin metabolite), but only occasionally objective tumour responses (Fjälling et al, 1996;McCarthy et al, 1998;Krenning et al, 1999). 111 In emits mainly photons that will cause a high whole body irradiation, which limits its use for therapy . 111 In also emits Auger electrons with extremely short range with the potential to cause cell death, if the radionuclide can be allocated to the nucleus (Bernhardt et al, 2003b). Using the highenergy beta emitter 90 Y, coupled to [Tyr 3 ]-octreotide via the chelator DOTA (1,4,7,10-tetraazacyclododecane,1,4,7,10-tetraacetic acid), partial tumour remission was achieved in 7 -24% of the patients (Paganelli et al, 2001;Waldherr et al, 2001;Valkema et al, 2002). Development of new somatostatin analogues with higher receptor affinity and higher degree of receptor internalisation will improve the therapeutic efficacy. The new somatostatin-based radioligand [DOTA 0 ,Tyr 3 ,Thr 8 ]-octreotide ( ¼ [DOTA 0 ,Tyr 3 ]octreotate) has a higher affinity for sstr2 than [DOTA 0 ,Tyr 3 ]octreotide, and coupled with the medium-energy b-emitter 177 Lu, high absorbed doses can be achieved (de Jong et al, 1997;Reubi et al, 2000;de Jong et al, 2001). 177 Lu has a shorter range than 90 Y and is therefore better suited for therapy of smaller tumours Bernhardt et al, 2003a). Previous experimental studies with xenografts have been performed using rat pancreatic acinar tumours carrying sstr, while the GOT1 model is an authentic human carcinoid with preserved phenotype (de Jong et al, 1998(de Jong et al, , 2001Kölby et al, 2001). In a recent series, patients with large NE tumour burden were treated with [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate and partial tumour regression was seen in 35% and complete remission in 3% (Kwekkeboom et al, 2003).
In the present study, we analysed the effect of treatment with [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate of the human midgut carcinoid GOT1 xenografted to nude mice. Our aims were to optimise the uptake of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate by defining the point of receptor saturation following a single i.v. injection of the radiopharmaceutical and to evaluate the effect of sstr-mediated radiation therapy with [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate given to xenografted midgut carcinoid tumours.

Animal model
The transplantable human midgut carcinoid GOT1 was established as previously described . In brief, cultured GOT1 cells were inoculated subcutaneously into nude mice. After 6 months, tumours became visible. These tumours were minced into pieces and new generations of tumours were generated by transplantation of tumour tissue to the subcutis of nude mice. When the tumours were 5 -10 mm in size, the experiments were started. The experiments were approved by the Ethical Committee for Animal Research at Göteborg University. All animal procedures were consistent with the animal use guidelines of the UKCCCR.
[DTPA]-octreotide was labelled with 111 In according to the manufacturer's instructions (Mallinckrodt Medical B.V., The Netherlands).
The peptide-bound fraction of 177 Lu and 111 In was assessed by instant thin layer chromatography (ITLC-SG, Gelman, Ann Arbor, MI, USA) with 0.1 M sodium citrate (pH ¼ 5.0, VWR International AB, Sweden) as mobile phase. The fraction of peptide-bound 177 Lu and 111 In was more than 99%.
The exact amount of radiopharmaceutical administered was determined by measuring the activity in the syringes before and after administration of the radiopharmaceutical with a well-type ionisation chamber (CRC-15R, Capintec, NJ, USA).

Biodistribution and biokinetics of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate
For studies of the biodistribution of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, five groups of tumour-bearing animals were injected with a single i.v. injection of 7.5, 15, 30, 60 and 120 MBq of 177 Lu-DOTA-Tyr 3 -octreotate (30 MBq mg À1 ), respectively (n ¼ 5 in all groups). The animals were killed 24 h later and tumour tissue and multiple organs were collected for measurement of radioactivity and histopathological analyses. The 177 Lu activity was measured using a Wallac 1480 gamma counter (WIZARD s 3 00 , Wallac, Oy, Finland). Correction was made for detector background and radioactive decay. The activity concentration (C) of the radionuclide was expressed as the fraction of injected activity per unit mass of the tissue (%IA g À1 ).
For studies of biokinetics, tumour-bearing animals were injected with 0.5 MBq (1 mg, n ¼ 20) or 30 MBq (1 mg, n ¼ 20) [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, that is, diagnostic and therapeutic amounts of radionuclide, respectively. All animals were injected with a single i.v. injection and were killed in groups of five at 1, 3, 7 and 13 days post injection (p.i.), respectively. Control animals (n ¼ 5) were untreated. Tumour tissue and multiple organs were collected for measurements of radionuclide uptake and histopathological analyses.
The activity concentration (%IA g À1 ) was determined for tumour tissue (C T ) and normal tissue (C N ). The tumour-to-normal tissue activity concentration ratio (TNC) was calculated according to the formula TNC N ¼ C T /C N .

Dosimetry
The calculation of the absorbed dose to tumours, kidneys, liver and bone marrow was based on the pharmacokinetics of 0.5 and 30 MBq [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate (Loevinger, 1988). The cumulative activity concentration, that is, the total number of radioactive decays per mass unit, was determined by estimating the area under the curve for the activity concentration in each organ vs time. The electron energy emitted per decay was assumed to be 147 keV (Sowby, 1983). The contribution from photons was neglected. Also, the contribution from organs other than the target organ was neglected.

Biodistribution of [ 111 In-DTPA]-octreotide
In order to make it possible to compare the uptake of radiolabelled somatostatin analogue with previous clinical studies, the binding of radiolabelled octreotide was investigated. Each animal (n ¼ 5) was injected i.v. with 4 MBq (0.2 mg) of [ 111 In-DTPA]-octreotide 24 h before killing. Tumours and blood samples were collected and weighed, and the 111 In activity in tumour and blood was measured using the gamma counter. The TNC Blood was calculated as described above (Forssell-Aronsson et al, 1995).

Morphological analysis
Morphological analysis was carried out on all tumours included in this study to verify potential therapeutic effects. Tumour tissue from all animals were harvested and fixed in 4% paraformaldehyde (PF) in PBS at pH 7.4 for 24 h. Specimens were subsequently Remission by sstr-mediated radiation therapy L Kö lby et al dehydrated and embedded in paraffin wax. Parallel sections were counterstained with haematoxylin (htx) and eosin for morphological analysis, and with Ladewig staining to facilitate the detection of fibrin thrombi in tumour vessels. In all animals given a therapeutic amount of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, microscopic examination of liver, kidneys and bone marrow was performed in order to reveal any morphological changes related to radiation. Deparaffinised sections were preincubated with 5% nonfat dry milk followed by incubation overnight with primary antibodies, listed in Table 1. Antiserum was diluted in PBS containing 1% BSA and 0.1% sodium azide. Bound antibodies were visualised by the indirect immunoperoxidase technique (EnVisiont þ , cat. no. K4000, Dakopatts, Glostrup, Denmark).

Apoptotic cell count
The number of apoptotic tumour cells was investigated in tumour tissue from the biokinetic study (animals injected with 30 MBq 177 Lu-octreotate and followed up to 13 days). Haematoxylineosin-stained sections were used to identify apoptotic tumour cells according to morphological criteria (condensed and fragmented nuclei and eosinophilic cytoplasma). Three high power fields (HPF; Â 40 objective) were selected from each tumour for counting of apoptotic cells. To obtain representative areas of the tumour, a direction from the centre to the periphery of the tumour was randomly selected. Photographs of each HPF were then systematically taken: (i) in the centre, (ii) half way to the periphery and (iii) in the periphery of the tumour, all along the selected line. The number of apoptotic cells was counted on the photographs and the average number of apoptotic cells per HPF calculated (apoptotic cell count).

Statistical analysis
Differences between groups regarding uptake of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, TNC value and apoptotic cell count were analysed by one-way analysis of variance. Logarithmic transformation was used for variance stabilisation. Compensation for multiple comparisons was performed using the Dunnett's method (Hochberg and Tamane, 1987).
Differences in biokinetics between 30 and 0.5 MBq were analysed by two-way analysis of variance followed by t-test at each time point. Logarithmic transformation was used for variance stabilisation. Compensation for mass significance was performed according to Bonferroni-Holm (Hochberg and Tamane, 1987).
For analysis of the therapeutic effect of [ 177 Lu-DOTA 0 -Tyr 3 ]octreotate, the number of CR in the treatment groups (7.5 -120 MBq) was compared to the number of CR in the control group using Fisher's exact test. Compensation for mass significance was performed according to Bonferroni-Holm.
The uptake of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate in tumours was similar for both groups of animals 24 h p.i. However, the biokinetics of the activity concentration differed significantly between the two groups. In animals given 0.5 MBq [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, which did not retard tumour growth, the activity concentration declined over time and was less than 50% at 7 days and less than 20% at 13 days p.i. compared to the initial concentration. In contrast, in animals given 30 MBq, the activity concentration remained high over time and the biokinetic curves differed significantly both at 7 days p.i. (P ¼ 0.00044) and 13 days p.i. (P ¼ 0.00081) (Figure 2). The tumour volumes at 7 days p.i. and 13 days p.i. for animals given 30 MBq were reduced to 1678.5 and 1174.2% (mean7s.e.m.) compared to the initial volumes.

Tumour-to-normal tissue activity concentration ratio
Tumour-to-normal tissue activity concentration ratios of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate were determined up to 13 days after a single injection of 30 MBq. In general, the TNC value increased compared to the situation at 24 h p.i., which indicates a favourable change over time for the radiation to the tumour in relation to the radiation received by liver, kidneys and bone marrow. The increase in TNC value at 7 days p.i., compared to 24 h p.i., was significant for the bone marrow (P ¼ 0.0031), liver (P ¼ 0.033) as well as for the kidneys (P ¼ 0.00015) (Figure 3).

Dosimetry
Tumours treated with 30 MBq received about 120 Gy, corresponding to 4.0 Gy MBq À1 , when the absorbed dose was calculated using biokinetic data for the therapeutic amount of 30 MBq. For normal    later. After 60 MBq, two out of six animals with CR remained free of recurrence until being killed 6 -7 months later. After treatment with 120 MBq, six out of 10 animals with CR remained free of recurrence until being killed 3 -4 months after the remission (Figure 4).

Induction of apoptosis and necrosis after treatment with [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate
In untreated controls, tumour cells grew in solid sheets with small variation in nuclear size and without necroses. The tumour cells were positive for the general NE-marker CgA, serotonin (5-HT) and Vesicular Mono Amine Transporters (VMAT1 and 2). In animals treated with 30 MBq [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, a significant increase in apoptotic cell count, compared to controls, was observed both at 1 day p.i. (P ¼ 0.0000030) and 3 days p.i. (P ¼ 0.00000059). In animals killed 3 days p.i., all five tumours had large confluent necroses and intercellular oedema. In animals, killed 7 and 13 days p.i., the number of tumour cells was clearly reduced as was the apoptotic cell count and the oedema and necroses were replaced by fibrosis. The typical appearance of a tumour 13 days p.i. was that of almost complete fibrosis and absence of tumour cells ( Figure 5). The apoptotic cell count at each time point is presented in Figure 5F.
In the therapy groups (7.5 -120 MBq), all tumours, or tumour residues, were morphologically examined at the end of the observation period. Gross examination of the small residual tumour tissue, found in animals with CR in response to [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, revealed brownish nodules, 1 -2 mm in diameter. The microscopic analysis demonstrated a thin rim of fibroblasts in the periphery and a central part consisting of crystalline structures surrounded by inflammatory cells including macrophages and giant cells (Figure 6). Specific staining for fibrin (Ladewig) did not show any signs of thrombosis in treated tumours.

Side effects of radiation
Morphologic analysis of the liver, kidneys and bone marrow after [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate therapy revealed two animals with reactive inflammatory response in the portal fields of the liver. These inflammatory infiltrates were composed both of CD 3positive T cells and CD 20-positive B cells and had a Ki-67 proliferative index of o1%. In one animal, the kidney contained a fibrous scar. The bone marrow of all animals was normal with preserved haematopoietic cells.

DISCUSSION
This study presents the first systematic optimisation of sstrmediated radiation therapy with [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate of a human midgut carcinoid xenografted to nude mice. The first part of the study demonstrated that saturation of sstr in the tumours after a single i.v injection of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate was evident between 0.5 mg (15 MBq) and 1 mg (30 MBq). It is important to define this level, since suboptimal therapeutic amounts will not lead to maximal uptake and maximal absorbed doses in the tumours. The uptake, and hence the absorbed dose, in the kidneys, liver and bone marrow increased proportionally to the amounts administered. Therefore, amounts in excess of the tumour saturation level will increase adverse effects of radiation therapy.
The second part of the study described the markedly different biokinetics of [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, given in diagnostic vs therapeutic amounts. The activity concentration in tumours after being given diagnostic amounts (0.5 MBq, 1 mg) was reduced over time. On the other hand, in tumours from animals given a therapeutic amount (30 MBq, 1 mg), the concentration remained high over time and significantly differed from the animals given a diagnostic amount both at 7 and 13 days p.i. The preserved high activity concentration in the therapy group was unexpected and significantly contributed to the high absorbed doses in the tumours, and thereby a very high efficacy of the radiation therapy. The difference in biokinetics can to a large extent be explained by tumour shrinkage, but other factors may also be of importance, for example, upregulation of sstr expression, radionuclide sequestration, increased intratumoural pressure and changed tumour blood flow initiated by the radiation.
The entrapment of radionuclide and altered biokinetics in the tumours after a therapeutic amount resulted in significantly increased TNC values over time. Therefore, choice of a radionuclide with half-life long enough to take advantage of this favourable change is essential when sstr-mediated radiation therapy is planned.
The third part of the study showed dose-dependent tumour responses. In animals treated with 7.5 and 15 MBq [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, the absorbed doses were low and CR was seen only occasionally. These tumours started to grow again after 2 weeks of reduced volume. In animals treated with 30 -120 MBq [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate, the absorbed doses were high; all, but one, of these animals showed CR. It is very unusual to obtain CR in animal tumour models using radionuclide therapy. Besides the present study and our previous study on [ 177 Lu-DOTA 0 -Tyr 3 ]octreotate therapy in the human small-cell lung cancer cell line NCI-H69 xenografted to nude mice, CR is very unusual (de Jong et al, 2001;Schmitt et al, 2004). For a long period, carcinoid Remission by sstr-mediated radiation therapy L Kö lby et al tumours were considered to be resistant to radiation therapy. However, a palliative role for both external radiation and 131 I-MIBG radiation therapy of NE tumours was proposed some 10 years ago (Samlowski et al, 1986;Kimmig, 1994). The present study clearly shows that human midgut carcinoid xenografts can be cured by adequate radiation therapy with minimal adverse effects. Sstr-mediated radiation therapy using [ 177 Lu-DOTA 0 -Tyr 3 ]octreotate in clinical series of patients with large tumour burden has not resulted in high rates of CR, most probably due to the administration of suboptimal amounts of radiopharmaceutical (Kwekkeboom et al, 2003). Our experiments show the importance of defining the amount of peptide required to achieve receptor saturation. This should, therefore, be included as part of the dose planning before therapy. Also correct choice of radionuclide and somatostatin analogue is of importance to obtain a high absorbed dose, which is a pre-requisite for a good tumour response Bernhardt et al, 2003a).
The present study not only showed a high cure rate of human midgut carcinoids but also revealed an unexpected entrapment of 177 Lu in tumour tissue during therapy. As recently shown in a patient with limited lymphoglandular spread of midgut carcinoid tumour, [ 177 Lu-DOTA 0 -Tyr 3 ]-octreotate has the potential to induce almost complete remission . GOT1 cells were harvested from the first patient, who underwent therapy with [ 111 In-DTPA]-octreotide. The biological characteristics of GOT1 closely resemble the tumour of the original patient, that is, preserved neuroendocrine differentiation and molecular markers,  preserved sstr expression and a TNC Blood value of about 400 for [ 111 In-DTPA]-octreotide uptake (Fjälling et al, 1996;Kölby et al, 2001). In other aspects, that is, subcutaneous tumour localisation and more rapid tumour growth, the model differs from the human situation. An orthotopic model, with liver metastases, would be very helpful to further resemble the clinical situation. The exceptional results obtained in GOT1 xenografts may thus indicate a possibility for more successful treatment of patients with metastatic carcinoid tumours than hitherto achieved in clinical series. In the individual patient, determination of sstr expression profiles in tumours by real-time quantitative PCR may be helpful to select the best somatostatin analogue to obtain maximal uptake of the radiopharmaceutical. Also, a radionuclide with suitable range in relation to the volumes of the tumours intended to be treated is important . Determination of the specific uptake in tumours (studied in biopsies or determined by octreotide scintigraphy) can also be helpful to estimate the absorbed dose and expected biological effect (Forssell-Aronsson et al, 1995). In the future, the dose dependency of the shown entrapment as well as the tolerance of normal tissues to sstrmediated radiation therapy must be studied in detail before this type of treatment can be transferred to the clinical situation with high success rate.