PEGylation, increasing specific activity and multiple dosing as strategies to improve the risk-benefit profile of targeted radionuclide therapy with 177Lu-DOTA-bombesin analogues

Background Radiolabelled bombesin (BN) conjugates are promising radiotracers for imaging and therapy of breast and prostate tumours, in which BN2/gastrin-releasing peptide receptors are overexpressed. We describe the influence of the specific activity of a 177Lu-DOTA-PEG5k-Lys-B analogue on its therapeutic efficacy and compare it with its non-PEGylated counterpart. Methods Derivatisation of a stabilised DOTA-BN(7–14)[Cha13,Nle14] analogue with a linear PEG molecule of 5 kDa (PEG5k) was performed by PEGylation of the ϵ-amino group of a β3hLys-βAla-βAla spacer between the BN sequence and the DOTA chelator. The non-PEGylated and the PEGylated analogues were radiolabelled with 177Lu. In vitro evaluation was performed in human prostate carcinoma PC-3 cells, and in vivo studies were carried out in nude mice bearing PC-3 tumour xenografts. Different specific activities of the PEGylated BN analogue and various dose regimens were evaluated concerning their therapeutic efficacy. Results The specificity and the binding affinity of the BN analogue for BN2/GRP receptors were only slightly reduced by PEGylation. In vitro binding kinetics of the PEGylated analogue was slower since steady-state condition was reached after 4 h. PEGylation improved the stability of BN conjugate in vitro in human plasma by a factor of 5.6. The non-PEGylated BN analogue showed favourable pharmacokinetics already, i.e. fast blood clearance and renal excretion, but PEGylation improved the in vivo behaviour further. One hour after injection, the tumour uptake of the PEG5k-BN derivative was higher compared with that of the non-PEGylated analogue (3.43 ± 0.63% vs. 1.88 ± 0.4% ID/g). Moreover, the increased tumour retention resulted in a twofold higher tumour accumulation at 24 h p.i., and increased tumour-to-non-target ratios (tumour-to-kidney, 0.6 vs. 0.4; tumour-to-liver, 8.8 vs. 5.9, 24 h p.i.). In the therapy study, both 177Lu-labelled BN analogues significantly inhibited tumour growth. The therapeutic efficacy was highest for the PEGylated derivative of high specific activity administered in two fractions (2 × 20 MBq = 40 MBq) at day 0 and day 7 (73% tumour growth inhibition, 3 weeks after therapy). Conclusions PEGylation and increasing the specific activity enhance the pharmacokinetic properties of a 177Lu-labelled BN-based radiopharmaceutical and provide a protocol for targeted radionuclide therapy with a beneficial anti-tumour effectiveness and a favourable risk-profile at the same time.


Background
Prostate and breast cancers are the most frequently diagnosed forms of cancer in the USA. Especially in addressing metastatic and small-volume diseases, it is essential to investigate, alongside conventional therapies, alternative treatments, such as peptide receptor radionuclide therapy (PRRT). The fact that certain tumour types overexpress, receptors for peptide-hormones provide the basis for successful use of radiolabelled peptide analogues as tumour tracers in nuclear medicine. The mammalian gastrin-releasing peptide receptor (BN 2 /GRP) [1,2] is particularly overexpressed in several human tumours, including prostate, breast and smallcell lung cancers [3][4][5]. The tetradecapeptide bombesin (BN) shows high binding affinity for these BN 2 /GRP receptors. Using BN conjugates for specific delivery of radionuclides into the above-mentioned tumours is therefore a promising strategy for diagnostic and therapeutic purposes.
BN analogues, however, present certain problems regarding therapy. They show poor enzymatic stability in vivo, which might prevent sufficient localisation at the target site. Furthermore, high accumulation and retention in healthy organs, which express the BN 2 /GRP receptor, increase the risk of side effects. Moreover, kidney toxicity, which was observed and investigated in PRRT with somatostatin analogues in clinical studies [6][7][8], may also hold true for BN analogues. Finally, several side effects were elicited from intravenous (i.v.) injection of BN agonists in humans. Therefore, a high specific activity of the radiolabelled BN agonist may be important in minimising such undesired effects.
Our preclinical study with a series of 99m Tc(CO) 3 labelled PEGylated BN analogues showed that PEGylation is an effective strategy to improve the therapy-relevant characteristics, which include higher tumour uptake, improved tumour retention and lower uptake into nontarget tissue. The PEG entity of 5 kDa was established as the optimal PEG size because it improved these features best [26].
The BN analogues of the current study were therefore based on one of our stabilised analogues (Gln 7 -Trp 8 -Ala 9 -Val 10 -Gly 11 -His 12 -Cha 13 -Nle 14 -NH 2 ) containing a β 3 hLys-βAla-βAla spacer ( Figure 1) [14]. The peptide was equipped with a 1,4,7,10-tetraazacyclododecane-1,4,7,10tetraacetic acid (DOTA) chelator to provide the analogue DOTA-β 3 hLys-βAla-βAla-Gln 7 -Trp 8 -Ala 9 -Val 10 -Gly 11 -His 12 -Cha 13 -Nle 14 -NH 2 (referred to as DOTA-Lys-BN, Figure 1a). We hypothesised that PEGylating this DOTA-Lys-BN would lead to the same favourable characteristics seen with PEGylated 99m Tc-based BN analogues. Derivatisation of the DOTA-Lys-BN analogue with a linear PEG molecule of 5 kDa (PEG 5k ) was performed by PEGylation of the E-amino group of the lysine residue. The resulting PEGylated BN (referred to as DOTA-PEG 5k -Lys-BN, Figure 1b) as well as the DOTA-Lys-BN were then radiolabelled with 177 Lu. We chose this radionuclide because it is currently used together with 90 Y for PRRT with somatostatin analogues on a routine basis in clinics [27,28] and because it proved to be less problematic concerning kidney toxicity in comparison with the 90 Y-radiolabelled somatostatin analogue [8,27]. Furthermore, application of 177 Lu allows imaging and PRRT at the same time owing to γ-ray emissions of suitable energy for SPECT, which enables dosimetry calculations and therapy monitoring [29].
In the current study, the new 177 Lu-labelled DOTA-Lys-BN and DOTA-PEG 5k -Lys-BN analogues were tested in vitro in human prostate carcinoma PC-3 cells and in PC-3 tumour bearing mice. They were compared in order to evaluate the effect of PEGylation on in vivo pharmacokinetics and their therapeutic effectiveness. Apart from looking at the anti-tumour efficacy, we also investigated the optimal risk-benefit profile by varying the specific activity of the radiolabelled DOTA-PEG 5k -Lys-BN analogue and assessed the efficacy of PRRT by varying the number and the interval of the 177 Lu-DOTA-PEG 5k -BN doses. For an estimation of potential kidney toxicity, the renal function was monitored with quantitative 99m Tc-DMSA scintigraphy.

Methods
Sources of materials, equipment, peptide synthesis and PEGylation are presented in Additional file 1.

Statistical analysis
All data are presented as mean ± SD. The in vivo data were statistically analysed with a t test (Microsoft Excel software). All analyses were 2-tailed and considered as type 3 (two-sample unequal variance); P < 0.05 was considered statistically significant.

Metabolic stability in human plasma
The labelled analogues were incubated with human plasma (final concentration, 10 MBq/0.6 ml) at 37°C for various time intervals up to 12 days. After incubation, proteins were precipitated with acetonitrile/ethanol (1:1) and TFA (0.1%) and then centrifuged. The supernatant was analysed with RP-high-performance liquid chromatography (HPLC) equipped with a radioactivity detector. The radioactivity chromatograms showed different peaks which corresponded to the intact peptide and the different degradation products. The experiments were performed two times.
For externalisation, PC-3 cells were incubated with the labelled analogues (60 kBq) in culture medium at 37°C for 1 h. After incubation, the supernatant was discarded, and the cells were twice washed with cold PBS. The cells were then incubated again at 37°C in culture medium for 0.5, 1, 2.5, 5 and 24 h. At each time point, the supernatant was collected, the cells twice washed with cold PBS and lysed with 1 N NaOH. The supernatant (released radioactivity) and the cells (bound/internalised radioactivity) were measured in the gamma counter. All experiments were carried out two to three times in triplicate.

Biodistribution studies
All animal experiments were conducted in compliance with the Swiss animal protection laws and with the ethical principles and guidelines for scientific animal experimentation established by the Swiss Academy of Natural Sciences. Biodistribution studies were performed with 6-to 8-week-old female CD-1 nu/nu mice (20 to 25 g) purchased from Charles River Laboratories (Sulzfeld, Germany). For the induction of tumour xenografts, each mouse received subcutaneously 8 × 10 6 PC-3 cells in 150 μl culture medium without supplements. The tumours were allowed to grow for at least 3 weeks. On the day of the experiment, the mice (3 to 6 per group) received the radioactive conjugates intravenously. For the biodistribution studies, the mice were injected with different specific activities of the radiolabelled BN analogues (low specific, 6.6 MBq/nmol peptide; high specific, 66 MBq/nmol peptide). Receptor-blocking studies were performed using 100 μg of unlabelled BN (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14) co-injected with the corresponding radiolabelled BN analogue. At 1, 4 and 24 h post injection (p.i.), the animals were euthanised and dissected. Blood, tumours and various healthy tissues and organs were collected, weighed and examined for radioactivity. Results are expressed as percentage of injected dose per gram of tissue (% ID/g).

Dose calculation
The absorbed doses to PC-3 tumours and critical organs were calculated from the biodistribution studies (1 MBq/0.1 ml; 0.3 or 3.0 nmol peptide; n = 3 per group). Under the assumption of rapid accumulation (uptake at 0 h p.i. corresponds to the uptake at 1 h p.i.), the cumulative radioactivity in each tissue was calculated (MBq/h) taking biologic elimination and physical decay into account up to 24 h p.i. and afterwards only physical decay up to 400 h p.i. The absorbed tumour doses of the mouse experiments were extrapolated from the sphere model doses which were calculated by using the software OLINDA (OLINDA/EXM1.0, Vanderbilt University, Nashville, TN, USA). The S values for all other tissues of mice were taken from E Larsson [30]. The absorbed dose (milligray per mega-Bequerel) was calculated by multiplying the area under the curve (AUC) (h; normalised to 1 MBq ID) with the S value (mGy/(MBq ? s)). The dose (in Gy) was calculated by multiplying the absorbed dose (mGy/MBq) with the amount of radioactivity injected (20 MBq). The AUC-estimate for an adult male was obtained by multiplying the AUC of the mice (MBq/h) with a factor consisting of (total body weight mouse /total body weight adult male ) × organ weight adult male . The subsequent dose calculation was performed using the adult male model of the software OLINDA.

Therapy studies
Therapy studies were conducted in mice bearing PC-3 xenografts. The tumour was assumed to be an ellipsoid, and its volume was calculated with the formula V t ¼ π=6 ð ÞLW 2 where L represents the longest dimension and W the shortest dimension of the tumour.
Two weeks post PC-3 inoculation, i.e. the day of the first injection (day 0), the tumours had an average volume of 85 mm 3 . The animals were divided into six groups of six mice ( Table 1). The control group received an i.v. injection of PBS only (group A). Another group was injected with unlabelled DOTA-PEG 5k -Lys-BN at a peptide amount of 3.0 nmol (group B). The treated mice received two equal doses of 20 MBq i.v. either of 177 Lu-DOTA-PEG 5k -Lys-BN (groups C to E) or of 177 Lu-DOTA-Lys-BN (group F) at a peptide amount of 0.3 or 3.0 nmol. The injections were performed either at days 0 and 14 or at days 0 and 7 (Table 1). Body weight and tumour volume of all mice were quantified 3 times per week. The relative volume of tumours V r was defined as V r = V t /V 0 , where V t was the measurement at time t (days after the first injection), and V 0 was the measurement at day 0. If a tumour disappeared, V t was set to 0. Mice were removed from the study promptly upon fulfilling one or both of the following criteria: a tumour volume of ≥ 1.5 cm 2 or total body weight loss of ≥ 15%. Upon euthanasia, tumours were collected and embedded in TissueTek (Sakura Finetek, USA Inc., Torrance, CA, USA) and frozen for autoradiography.
99m Tc-DMSA SPECT/CT imaging studies Three groups of four mice each (groups G to I, Table 1), which were not xenografted with PC-3 cells, were included in the therapy study for 99m Tc-DMSA tests. An untreated control group of mice (group G), a treated group receiving two doses of the 177 Lu-DOTA-PEG 5k -Lys-BN analogue at high specific activity (group H) and a treated group of mice getting two doses of the 177 Lu-DOTA-PEG 5k -Lys-BN analogue with low specific activity (group I). 99m Tc-DMSA scans with SPECT/computed tomography (CT) were obtained 43, 71 and 111 days after therapy, 2 h after i.v. injection of about 30 MBq 99m Tc-DMSA. SPECT scans were acquired with anaesthetised mice during 20 min using 15 projections/min. The images were obtained on an X-SPECT-system (Gamma Medica, Inc., Northridge, CA, USA) equipped with a single head SPECT device and a CT device. SPECT data were acquired and reconstructed with LumaGEM (version 5.407, Gamma Medica, Northridge, CA, USA). CT data were acquired with an X-ray CT-system (Gamma Medica) and reconstructed with the software CoBRA (

In vitro evaluation
The PEGylation of the DOTA-Lys-BN analogue (Additional file 1: Figure S7) as well as the results of the log D and IC 50 determination are presented in Additional file 1. PEGylation resulted in a slightly increased hydrophilicity of the analogue and in an eightfold higher IC 50 value compared with that of the non-PEGylated analogue ( Lu-DOTA-Lys-BN was rapidly degraded by proteases in human plasma. After 5 days of incubation, it was almost entirely metabolised, and only 13.8 ± 5.7% remained intact. PEGylation resulted in a marked increase in protease stability; 51.8 ± 0.9% of 177 Lu-DOTA-PEG 5k -Lys-BN still remained intact after 5 days of incubation and 43.7 ± 0.5% after 11 days. Moreover, the half-life of 177 Lu-DOTA-Lys-BN in human plasma was 1.2 ± 0.3 days, whereas the half-life of 177 Lu-DOTA-PEG 5k -Lys-BN was 6.7 ± 1.4 days ( Figure 2). 177 Lu-DOTA-Lys-BN internalised rapidly into PC-3 cells and reached its maximum within the first hour of incubation (approximately 30%/10 6 cells). The PEGylated analogue showed a significantly lower and slower internalisation into PC-3 cells. After incubation for 4 h, the internalised fraction was 3.3 ± 1.2%. Externalisation studies revealed 63.1 ± 4.0% of the internalised 177 Lu-DOTA-Lys-BN externalised within the first 2.5 h. After 24 h, only 13.5 ± 7.2% of the internalised fraction was found in the cells. In contrast, the externalisation of the PEGylated analogue was slower (24.2 ± 1.3% after 24 h).
The tumour-to-non-target ratios were similar for both derivatives. The initial longer circulation time of 177 Lu-DOTA-PEG 5k -Lys-BN, however, resulted in lower tumour-to-blood ratios at 1 and 4 h p.i. compared with that of the non-PEGylated compound. 177 Lu-DOTA-PEG 5k -Lys-BN showed a twofold increase in the tumour-to-pancreas ratio at all time points and in tumour-to-kidney and tumour-to-liver ratios at 24 h p.i. (Figure 3).
In additional biodistribution studies, two ligand doses of the 177 Lu-DOTA-PEG 5k -Lys-BN at a peptide amount corresponding to the amount that was injected in the therapy studies (high specific, 0.3 nmol; or low specific, 3.0 nmol peptide injected per mouse) were administered. This showed that the uptake in the receptor-expressing tissues such as the pancreas and colon was markedly reduced by applying a high amount of PEGylated peptide (3.0 nmol). The tumour uptake was also reduced by 29% and 42% at 1 h p.i. and 24 h p.i. after injection of a high amount of peptide ( Figure 4, Table 3). In comparison, the tumour-to-blood, tumour-to-liver, tumour-to-kidney and tumour-to-muscle ratios were approximately twofold higher at all time points if 177 Lu-DOTA-PEG 5k -Lys-BN was injected at a low molar amount of peptide (0.3 nmol). The only ratios which revealed a higher value with a high amount of peptide (3.0 nmol) were the tumour-to-pancreas ratios (Table 3).

Therapy studies
The therapy study was performed according to the protocol shown in Table 1. In total, 48 mice were included and divided into six groups of mice bearing PC-3 tumours (groups A to F; n = 6) and three additional groups of mice without tumours (groups G to I; n = 4). All groups of mice which received the 177 Lu-labelled BN analogues (C to F) clearly showed a reduction of tumour growth in comparison with groups A and B which received only PBS or unlabelled BN analogue. The treatment with the non-PEGylated 177 Lu-labelled BN analogue (0.3 nmol; group F) significantly decreased the PC-3 tumour growth rate with respect to that of group A and exhibited an inhibition of 53% 3 weeks after the first dose ( Figures 5 and 6). The tumour growth inhibition was higher (63%) with the 177 Lu-DOTA-PEG 5k -Lys-BN analogue of high specific activity (group D). However, the 177 Lu-DOTA-PEG 5k -Lys-BN analogue of low specific activity (group C) exhibited only an inhibition of 36%. The most effective tumour growth inhibition of 73% (3 weeks after the first dose) was observed when the second dose of the PEGylated BN analogue of high specific activity was applied 7 days after the first dose (group E; Figure 6). Mice of group B did not show an increased tumour growth with respect to group A ( Figure 5), although BN agonists are known to have mitogenic characteristics.
99m Tc-DMSA SPECT/CT imaging studies Forty-three days after therapy, the renal 99m Tc-DMSA uptake of the treated animals (group H) receiving the radiotracer of high specific activity was 76,397 counts/ kidney, whereas the uptake of the treated animals receiving the radiotracer of low specific activity (group I) was 74,949 counts/kidney. Seventy-one days after therapy, there was no significant difference in the renal 99m Tc-DMSA uptake between groups G, H and I (51,344, 57,147 and 47,692 counts/kidney, respectively); 111 days after therapy, there was also no significant difference in the renal 99m Tc-DMSA uptake between these three groups of mice.
In vitro, time-dependent cell uptake and internalisation showed slower binding kinetics for the PEGylated BN analogue. These findings are in line with the results of PEGylating other biomolecules reported in the literature [31]. PEG is also reported to affect target association and dissociation rates of antibody fragments negatively [32]. These aspects may apply to our 177 Lu-DOTA-PEG 5k -Lys-BN and explain why binding affinity of this analogue in vitro was slightly reduced (Additional file 1), the steady state was reached later, and the total cell binding was lower in comparison with that of the non-PEGylated counterpart.
The biodistribution data, in which 0.002 nmol of 177 Lu-AMBA and 177 Lu-DOTA-8-AOC-BN(7-14) (HPLC purified) was injected per mouse [23], and the data of 177 Lu-DOTA-PESIN (0.2 nmol peptide) [25] were compared with our biodistribution data, in which 0.075 nmol of the 177 Lu-labelled BN analogues were injected. This 0.075 nmol is the nearest possible approximation to the 0.002 nmol without HPLC purification, which is desired in clinics. In comparison with 177 Lu-AMBA, our 177 Lu-DOTA-Lys-BN showed an approximately fourfold lower kidney uptake 1 h p.i., whereas the kidney uptake of the 177 Lu-DOTA-PEG 5k -Lys-BN analogue was 2.3-fold lower at 1 h p.i. Both compounds showed a faster clearance from the kidneys within 24 h p.i. Kidney accumulation and washout of our 177 Lu-DOTA-Lys-BN and 177 Lu-DOTA-PEG 5k -Lys-BN were comparable to those of 177 Lu-DOTA-PESIN (3.8 ± 0.34% ID/g at 1 h p.i.), even though Gelofusine and polyglutamic acid were co-administered with 177 Lu-DOTA-PESIN for the reduction of renal uptake [25]. Furthermore, the GI uptake was much lower with 177 Lu-DOTA-Lys-BN and 177 Lu-DOTA-PEG 5k -Lys-BN at 1 and 24 h p.i. compared with that in 177 Lu-AMBA (11.2%ID and 5.8% ID, respectively) and 177 Lu-DOTA-8-AOC-BN(7-14) (9.7% ID and 1.7% ID, respectively) [23]. However, the significantly higher blood level at 1 h p.i. after PEGylation might cause higher bone marrow toxicity and could therefore be a potential drawback of PEGylation. 177 Lu-DOTA-PEG 5k -Lys-BN showed significantly higher tumour uptake at 1 h p.i. in comparison with the non-PEGylated counterpart. The higher enzymatic stability as well as the longer blood circulation may have compensated for the slower binding kinetics and the lower receptor affinity of DOTA-PEG 5k -Lys-BN. In order to compare the cumulative radioactivity over 24 h of each conjugate in the tumour, the AUC value of 177 Lu-DOTA-Lys-BN was arbitrarily set to 1. The comparison showed a relative AUC value of 1.6 (P < 0.0006) for 177 Lu-DOTA-PEG 5k -Lys-BN.
The second hypothesis that PEGylation prolongs the tumour retention was also proven. Even though PEGylation lowered the tumour washout only slightly between 1 and 24 h p.i., there was more 177 Lu-DOTA-PEG 5k -Lys-BN retained in the tumour between 0 and 24 h p.i. The extended tumour retention for the 177 Lu-DOTA-PEG 5k -Lys-BN might be explained by the improved enzymatic stability of the peptide derivative, and the extended retention might be due to the enhanced permeation and retention in the tumour. On the basis of the biodistribution data with 177 Lu-AMBA [23] and 177 Lu-DOTA-PESIN [25], both BN analogues showed higher tumour uptakes (6.35 ± 2.23% ID/g and 11.6 ± 1.4%ID/g at 1 h p.i., respectively) and better retention profiles than our 177 Lu-labelled BN analogues. However, compared to 177 Lu-DOTA-8-AOC-BN (7)(8)(9)(10)(11)(12)(13)(14) [23] (2.84 ± 1.65% ID/g at 1 h p.i.), our 177 Lu-DOTA-Lys-BN analogue showed a similar tumour uptake, but the uptake of 177 Lu-DOTA-PEG 5k -Lys-BN was higher. This comparison, however, must be looked at with due care because the study designs differ insofar as different peptide amounts were injected.
The third hypothesis, i.e. that PEGylation improves tumour-to-non-target ratios, could partially be confirmed. The tumour-to-non-target ratios were rather similar for both derivatives. However, in comparison with the non-PEGylated BN analogue, the 177 Lu-DOTA-PEG 5k -Lys-BN analogue exhibited a higher tumour uptake and a prolonged tumour retention which resulted in increased tumour-to-pancreas ratios at all time points and in higher tumour-to-liver and tumour-to-kidney ratios at 24 h p.i. (Figure 3).
Alongside PEGylation, the influence of the specific activity on biodistribution was evaluated. 177 Lu-DOTA-PEG 5k -Lys-BN injected at two different peptide amounts corresponding to the amount that was injected in the therapy studies (0.3 or 3.0 nmol, respectively) affected the uptake into receptor-expressing tissues. The amount of 0.3 nmol was selected to approximate the 0.22 nmol of the AMBA therapy study because these amounts of 0.3 nmol have proven to be the limit for high specific labelling, i.e. the labelling is reproducible without any loss in yield. The amount of 3.0 nmol however was selected because a preliminary study (data not presented) had suggested that peptide amounts in this range markedly reduce the uptake into non-target receptor positive tissues. In comparison with a low peptide amount, applying a high peptide amount resulted in a marked reduction in pancreas and colon uptake which would lower the risk of radiotoxic side effects induced by radionuclide therapy (Figure 4). However, the cumulative radioactivity in the tumour was significantly higher with a low peptide amount. The dosimetry showed that the absorbed dose into the tumour was 1.7-fold higher with the radiotracer of high specific activity, which would presumably indicate a higher anti-tumour effect. Furthermore, a lower accumulation in the kidneys within 24 h p.i. was observed with 177 Lu-DOTA-PEG 5k -Lys-BN at a low amount of peptide (Table 3), which would indicate a reduced risk of nephrotoxicity induced by radionuclide therapy. Thus, the incidence of BN-related toxicity after i.v. injection could be reduced using a low amount of peptide.
The radionuclide therapy studies (Table 1) showed a higher anti-tumour effectiveness with 177 Lu-DOTA-PEG 5k -Lys-BN (group D) compared with 177 Lu-DOTA-Lys-BN (group F) (63% vs. 53% inhibition 3 weeks after the first dose, respectively; Figure 6). This is in accordance with the biodistribution data, which showed a higher tumour uptake and retention after PEGylation ( Table 2). As comparative time point, we chose 3 weeks after the first dose, in order to evaluate the effectiveness of the different therapy protocols. This is the latest time point before several mice had to be euthanised upon fulfilling the endpoint criteria. Therefore, an interpretation after 3 weeks is not reliable since the groups represent only individual mice ( Figures 5 and 6).
The therapy studies, in which the specific activity was varied (group C vs. group D), resulted in a markedly higher therapeutic efficiency when 177 Lu-DOTA-PEG 5k -Lys-BN was applied at high specific activity (63% vs. 36% inhibition 3 weeks after the first dose). The lower tumour accumulation of 177 Lu-DOTA-PEG 5k -Lys-BN of low specific activity resulted in a proportionally faster tumour growth. We could demonstrate that the reduced efficacy is not caused by the tumour growth-promoting effect of the higher peptide amount since unlabelled DOTA-PEG 5k -Lys-BN (group B) did not induce tumour growth compared with the control group ( Figure 5). These results are in line with previous observations reported in the literature [25]. The high specific therapy, as we have seen, was more efficient than the low specific, which is in accordance with the biodistribution studies which demonstrate that the uptake in GRPRexpressing tissues is highest for the lower peptide dose and is reduced with the higher peptide dose. This phenomenon is considered to be the result of partial saturation of receptors in the target tissues at higher peptide doses.
Furthermore, it can be assumed that an increase in specific activity would achieve at least the same therapeutic efficiency as low specific activity, but the dosage injected would be lower.
Preliminary therapy studies (Additional file 1: Figure  S8), as expected, showed that the administration of two doses (2 × 20 MBq = 40 MBq) was more effective in tumour growth inhibition than application of a single dose (20 MBq). As shown with in vitro autoradiography (Additional file 1: Figure S9), there was no long-lasting down-regulation of BN 2 /GRP receptors in the tumour after treatment, which suggests that it is sensible to apply a multiple dosage. Therefore, two different twodose regimens were evaluated in the current therapy studies. Applying the second dose at day 14 was chosen to match the AMBA therapy study. The preliminary study showed that the tumour started to grow after 14 days regardless of the second injection. Since the cause for this might have been that the tumour was already too large to respond to the treatment, the second application was introduced at day 7 in order to hit the tumour in an earlier state. The therapeutic efficiency was increased even further when the second dose of 177 Lu-DOTA-PEG 5k -Lys-BN (group E) was applied 7 days after the first dose instead of 14 days (73% vs. 63% inhibition at day 21) (group D).
In order to assess the risk for nephrotoxicity related to radionuclide therapy, a rough dosimetric estimate for an adult male was performed based on the biodistribution, in which 0.3 or 3.0 nmol of the 177 Lu-DOTA-PEG 5k -Lys-BN analogue were applied. This estimate implies that an administration of approximately 2 GBq of either low or high specific 177 Lu-DOTA-PEG 5k -Lys-BN analogue would result in absorbed kidney doses of approximately 18.8 or 24.6 Gy, respectively. These doses would not exceed the acceptable safe limit of 23 to 27 Gy [33]. The administration of 2 GBq would supposedly be necessary to reach a tumour dose of 50 Gy (supposed that the absorbed dose into the pancreas corresponds to the tumour dose), which is needed for treatment as external beam radiation therapy and brachytherapy data suggest [34][35][36]. A further step in the risk assessment was 99m Tc-DMSA scintigraphy which showed that there was no kidney damage in the mice treated with high or low specific 177 Lu-DOTA-PEG 5k -Lys-BN analogue (group H and I) since there was no significant difference in renal 99m Tc-DMSA uptake of control and treated mice. Besides, serum analysis confirmed the absence of renal toxicity (Additional file 1).

Conclusions
PEGylation, increasing the specific activity of the radiolabelled bombesin analogue and shortening the injection interval proved to be effective strategies to enhance the radiotherapeutic efficacy and to provide a favourable risk-profile at the same time. Tumour targeting