Targeted chemotherapy using a cytotoxic somatostatin conjugate to inhibit tumor growth and metastasis in nude mice.

The major problems of traditional chemotherapy are non-selectivity and non-specificity, resulting in severe toxic side effects. Peptides are a new-generation of drug-delivery vector to increase efficacy of this therapy and avoid the resulting damage. The cytotoxic somatostatin (SST) conjugate JF-10-81 was developed by coupling camptothecin (CPT) to the N-terminus of a SST analog (JF-07-69) using an activated carbamate linker. This conjugate selectively targets somatostatin receptor subtype 2 (SSTR2) and also retains high binding affinity and rapid internalization as well as anti-proliferative activity towards various tumor cells. JF-10-81 was tested for its inhibitory activity against the growth of human tumors which included neuroblastoma (IMR32), pancreatic cancer (CFPAC-1), leukemia (MOLT-4), pancreatic carcinoid (BON) and prostate cancer (PC-3). Both SSTR2 mRNAs and proteins were detected in all these tumor cell lines. The conjugate displayed potent in vivo inhibitory activity, although some of the potency measured in in vitro experiments was lost. JF-10-81 was found to significantly inhibit growth of these SSTR-positive tumors, resulting in 87% tumor reduction in neuroblastoma IMR32 and 97% in leukemia MOLT-4 bearing animals, even inducing regression of CFPAC-1 tumors. SSTR-overexpressing BON tumors were unfortunately relatively CPT-insensitive in vitro, however, JF-10-81 again exhibited in vivo potency presumably by specifically increasing CPT concentrations inside the tumor cells so that the inhibition rate for JF-10-81 was 85%. Also, JF-10-81 was used to treat highly invasive PC-3 tumors where s.c. injections inhibited both tumor growth (almost 60% reduction) and tumor metastasis (over 70%). This conjugate demonstrated its broad and excellent anti-tumor activity by targeting SSTR2-specific tumor tissues, supporting that short peptides and their analogs may be applied as ideal drug-delivery carriers to improve the traditional chemotherapy.


Introduction
It is hard to overemphasize the myriad drawbacks to traditional chemotherapy used in treating advanced or metastatic cancers. However, the major problems of non-selectivity towards tumor cells and nonspecifi city certainly rank among the likely potential causes of severe toxic side effects. Also, the instability and/or insolubility of cytotoxic agents in aqueous media limit their clinical application. Possible routes to overcome these problems and improve their therapeutic properties are numerous and include various new drug delivery systems such as polymer [1,2], nano-particles [3][4][5][6], antibody conjugates [7,8], and short peptide hormone conjugates [9][10][11][12][13].
Specifi c receptors for many of these peptides, the most investigated being somatostatin (SST) and bombesin (BM), have been found to be expressed aberrantly in various tumors [14]. Taking advantage of these observations, receptor-targeted cytotoxic peptide conjugates can be used to target receptorspecifi c tumor tissues, increase effi cacy of chemotherapeutic agents and decrease severe toxic side effects as long as suitable cleavable linker groups can be synthesized. Hence, several peptide conjugate candidates have been developed during recent years and short, easily synthesized peptides [13] were used as vectors to couple with chemotherapeutic agents and carry them to target receptor-specifi c tumor cells. In our laboratories, camptothecin (CPT), a potent anti-tumor agent [15,16], has been successfully coupled to urotensin, BM, and SST analogs. These new cytotoxic conjugates displayed relatively low toxic side effects and signifi cant inhibition to tumor cell proliferation and tumor growth [11,12,13,17]. Also, cytotoxic conjugates of SST [18], BM [19] and luteinizing hormone-releasing hormone (LHRH) [20] coupled with 2-pyrrolinodoxorubicin, have been synthesized and investigated by others. These conjugates showed broad and excellent anti-tumor activities [18][19][20][21]. Peptide hormones as drug-delivering vectors have their own specifi c advantages over other approaches such as rapid internalization (via their respective GPCRs), better tumor-penetrating ability and quick circulatory clearance [14].
SSTRs, especially SSTR2, have been found to be highly expressed in various tumors [14,22,23] and tumoral blood vessels [24]. The two naturally occurring compounds, SST-14 and SST-28, bind with high affi nity to all fi ve receptor subtypes whereas certain clinically approved SST analogs only selectively bind to SSTR2 while being much more stable in the circulation. Such analogs are attractive starting points for use as vectors to precisely deliver non-selective cytotoxic agents to receptor-selective target sites. For instance, CPT coupled to a SST analog selectively bound to SSTR2 and was designed to retain high binding affi nity even with the attachment of large drug molecules. This CPT-SSA conjugate JF-10-81 has already shown favorable anti-invasive and antiangiogenic activities [25]. In the current study, we further extend these studies to investigate the ability of JF-10-81 to prevent the growth of several important tumor types by sustained delivery formulations and metastasis of the highly invasive prostate cancer PC-3 tumor by periodic injections.

Cell culture
The CHO-R1 and CHO-R2 cell lines, which were transfected with rat SSTR1 and SSTR2, respectively, were a kind gift from Dr. Agnes Schonbrunn (University of Texas-Houston, Houston, TX). Cells were cultured in F-12 medium containing 10% fetal calf serum. All other human tumor cell lines, including neuroblastoma IMR32, leukemia MOLT-4, pancreatic cancer CFPAC-1, prostate cancer PC-3 were purchased from ATCC (American Type Culture Collection, Manassas, VA). Human pancreatic carcinoid BON cells were a kind gift from Dr. Courtney Townsend (University of Texas-Galveston) and cultured as described previously [9,26].

Ligand-receptor binding and internalization assay
The assay was performed as a modifi cation of the one previously described [27,28]. The binding buffer contained Dulbecco's Modifi ed Eagle's Medium (DMEM), 50 mM HEPES, 2 mM glutamine, 11.5 mM glucose, 1% bovine serum albumin (BSA) and 5 mM sodium pyruvate. The binding reactions took place in 1.5 ml eppendorf tubes containing 70-100 pM of [ 125 I]Tyr 11 -SST-14 from PerkinElmer (Boston, MA 02118-2512), 1−1.5 × 10 5 cells and the tested SST analogs (10 −5 to 10 −13 M) for a total volume of 300 µl. These were incubated at room temperature for 50-60 minutes and centrifuged for 1-2 min at 9,000 g. The cell pellets were washed twice with phosphate-buffered saline (PBS) and counted for total radioactivity in a Titertek gamma counter (Huntsville, AL). The analyses were done using GraphPad Prism 4.0 (San Diego, LA). For further internalization [29], cells were then re-suspended and incubated in HBSS (pH4, acetic acid) on ice for 10 min, washed with cold PBS and then re-counted for radioactivity. Due to the radioactivity of membrane-bound receptors being released by the acid wash, the remaining radioactivity represents the internalized peptide.

Western blot
The protocol was employed as described (Santa Cruz, CA). Briefly, cells were harvested, resuspended in RIPA buffer with cocktail inhibitors, homogenized by passing through a 21 gauge needle, mixed with loading buffer containing fresh DTT and heated for 5 min at 95 o C. Supernatants were loaded to run on 8%-16% Tris-glycine gel after centrifugation at 10,000 g. Protein was transferred from gel to nitrocellulose membrane which was then blocked with 5% fat-free milk, washed and incubated with SSTR2 antibody (Santa Cruz). The membrane was washed again and incubated with second antibody (Santa Cruz). Eventually, fi lms were developed according to the ECL system protocol (Amersham Biosciences, England).

Cell proliferation assay
The kits for the cell proliferation assay (MTT Assay) were purchased from Promega (Madison, WI) and the protocol was the same as described previously [26].

In vivo experiments
Male nude mice, 5-7 weeks of age upon arrival, were purchased from the National Cancer Institute (NCI, Frederik, MD). The animal study was approved by the Tulane University Health Sciences Center Institutional Animal Care and Use Committee (IACUC).

Tumor growth
The tumor tissue segments, grown from CFPAC-1, IMR32, MOLT-4 and BON cells injected into nude mice, were s.c. implanted into the fl anks of nude mice as described previously [9]. All the tumors were passaged from mouse to mouse. Tumor-carrying mice then were implanted with continuous drugreleasing pellets containing 5 mg of JF-10-81.
For treatments of pancreatic cancer CFPAC-1 tumors, mice were separated into 2 groups. Each mouse in the control group was implanted with a placebo pellet. Each mouse in the test group was implanted with a pellet containing JF-10-81 (5 mg/ pellet) with a drug-release time of 60 days. Tumor volumes were measured and bodyweights were taken once a week. The treatment for all other tumors (BON, IMR32 and MOLT-4) was the same as above.

Tumor metastasis
The experiment was performed as described [31] with modifi cation. Human prostate cancer PC-3 cells were re-suspended in a PBS solution and further mixed with matrigel at a ratio of 1:3. The mixtures were implanted via i.p. injection at a volume of 200 ul (1 × 10 6 cell/ml). One week later, mice were divided into 3 groups of 8-9 mice each. The control group was treated with PBS. The test groups were treated with the cytotoxic conjugate JF-10-81 (3.5 mg/kg) by i.p. or s.c. injection, 3 times a week, for a total of 12 injections. Mice were monitored and checked each week. Eventually, mice were killed by over-anaesthetization and autopsies were performed and tumor metastasis investigated.

Results
High binding affi nity of the CPT-SST conjugate to SSTR subtype 2 Both native SST-14 (DC-44-76) acting as the control peptide and the CPT-SST conjugate JF-10-81 were tested. As expected in CHO-R1 and CHO-R2 cells, synthetic SST-14 exhibited high affi nity for SSTR1 (EC50: 1.85 nM) and SSTR2 (1.45 nM) ( Fig. 1), no affi nity to the non-transfected CHO cells themselves. The vector JF-07-69 (2.35 nM) and its cytotoxic conjugate JF-10-81 (2.46 nM) displayed similar high affi nity binding to SSTR2. However, both of them had relatively weak binding affi nity to SSTR1, the EC 50 values being 100 fold less than for SSTR2 (Fig. 1). The further internalization assay showed that the conjugate JF-10-81 could quickly internalize into the SSTR2-specifi c cells (data not shown). Therefore, this conjugate could potentially be applied for receptor-targeted chemotherapy.

Expression of SSTR2 in cultured tumor cells
It has been demonstrated that SSTRs, especially SSTR2, are aberrantly expressed in many tumors [14]. Herein, the expression of mRNAs of all 5 SSTRs and protein of SSTR2 in 5 tested tumor cells were investigated by RT-PCR and western blot.
As shown in Figure 2, the results of RT-PCR demonstrated that SSTR2 and SSTR5 mRNAs existed abundantly in all 5 tumor cell lines, except for a lack of SSTR5 in MOLT-4 cells. SSTR1 was highly expressed in BON and CFPAC-1 cells. CFPAC-1 cells showed the existence of mRNAs of all 5 SSTR subtypes, with traces of one or more subtypes detectable in the other cells (Fig. 2). Further, SSTR2 proteins (∼64 kD) were detected in all 5 tumor cells by western blot (Fig. 3). The expression of SSTR2 in all tested tumor cells was identical to previous reports that SSTR2 is abberantly expressed in these tumor tissues, thus confi rming that SSTR2 could be used as a specifi c target for JF-10-81 to treat these receptor-positive tumors.
Although not easily detected in many in vitro cultured tumor cells, binding affi nities in 5 tumor cells were further investigated with SST-14 which has high affi nity to all 5 SSTR subtypes. We found that BON

Inhibition of the CPT-SSA conjugate to in vitro tumor cell proliferation
The CPT-SSA conjugate (JF-10-81) has been shown to have anti-angiogenic activity via a series of in vitro and in vivo experiments [25]. In the  Table 1). All the tumor cells were relatively CPT-sensitive compared to CPT-insensitive BON cells. Also, JF-10-81 exhibited the same trends as CPT itself. However, generally the conjugate partly lost its cytotoxic activity, with its EC50 being almost 10-20 fold less when compared to CPT itself. As expected, the vector peptide JF-07-69 did not work even at high concentrations ( Fig. 4 and Table 1).
Inhibition of the CPT-SSA conjugate to tumor growth and metastasis JF-10-81 was further evaluated for its ability to inhibit tumor growth and metastasis in these SSTR2-positive tumor models mentioned above. PC-3 tumors were treated with single injections of the conjugate. The others were implanted with continuous drug-releasing pellets (5 mg/pellet). As shown in Table 2 and Figures 5 and 6, JF-10-81 (5 mg/pellet) displayed dramatic inhibitory activity in all tested tumors.
Notably, in the SSTR-over-expressing, but CPTresistant pancreatic BON tumors, JF-10-81 (5 mg/ pellet) appeared to display a much more potent inhibitory activity than did CPT alone (5 mg/pellet).  The percentage increase of tumor volumes was 755% in the control group, 71% in JF-10-81treated group (5 mg/pellet) ( Table 2) and 278% in the CPT-treated group (5 mg/pellet) (data not shown). The inhibitory rates for JF-10-81 and CPT treatment are 85% and 42%, respectively. On average, JF-10-81 clearly displayed potent anti-tumor activity (depending on cell type). This could perhaps be because this conjugate concentrated CPT inside tumor cells by targeting receptors on cell surfaces, although favorable changes in the metabolism and clearance of biologically active, conjugated CPT cannot be ruled out. Even at a lower dose (1 mg/pellet), JF-10-81 was effi cacious against pancreatic cancer CFPAC-1 tumor growth with the inhibitory rate reaching 56% . However, JF-10-81 at the same dose (1mg/pellet) did not work well in IMR32, MOLT-4 and BON tumor treatments (data not shown). These results also indicate that the inhibitory ability of JF-10-81 was dose-dependent.
The highly invasive PC-3 cells were chosen for anti-tumor metastasis experiments. PC-3 tumors in the control groups were found to metastasize to the gut, liver, chest, skin etc. of nude mice. However, in treated groups, we found that there was an obvious decrease in overall tumor metastasis, with no metastasis taking place in some mice. JF-10-81 at 3.5 mg/kg blocked almost 73% of tumor metastasis using s.c. injections and 60% by i.p. injections (Fig. 6). Meanwhile, JF-10-81 was observed to result in an almost 60% reduction of primary tumor weights using s.c. administration and 40% by i.p. injections.

Discussion
Peptide-based receptor-targeted chemotherapy is one of a new generation of promising approaches for increasing the specifi city of cancer-cytotoxin treatments. Peptides might be optimal drug-delivery vectors due to their unique advantages which have been discussed [14]. Certain peptides such as SST, BN, LHRH or their analogs have been used to deliver cytotoxic compounds by taking advantage of the peptide receptors being aberrantly expressed and sometimes over-expressed in many human tumors and in the blood vessels around tumors [13,14,22,23,24,32]. Furthermore, the structureactivity relationships of these peptides are well investigated and allow for attachment of large drug molecules using a large variety of cleavable linker chemistries with little or no loss of receptor affi nity and retention of signifi cant anti-tumor activities [9,10,20,21]. For example, AN-238, which was made by linking 2-pyrrolino-DOX to the SST  (5) 27.63 ± 0.88 (6) Ͼ10,000 Note: a Mean ± SEM (n) and the experiments were separately repeated n times.  analog RC-121, exhibited potent anti-tumor activity by targeting tumor cells [33,34] or tumoral blood vessels [10]. In our previous studies, cytotoxic conjugates of SST and BN analogs have revealed some utility against tumor growth and angiogenesis [9,25,35]. The CPT-SST conjugate JF-10-81 was found to be stable and readily water-soluble [9,25,36] and is predicted to display less systemic toxicity than the naked cytotoxic agent [36]. Also, JF-10-81 inhibited growth of SSTR-expressing tumors [9,11] and blocked development of angio-genesis [25]. In addition, a number of other peptide conjugates have been shown to have more specifi city to tumor cells with less toxicity to normal cells [12,13,37,38].
In the present work, our most widely studied conjugate (JF-10-81) was further tested for its potentially broad anti-tumor activity using human tumors ranging through neuroblastoma, pancreatic cancer, pancreatic carcinoid, leukemia and prostate cancer. With all these tumors, as expected, RT-PCR and western blot revealed the presence of SSTR2. The internalization assay demonstrated that this SSTR2-specifi c JF-10-81 could quickly be taken up in SSTR2-transfected cells. In our experience, a frequent low dose was found to work better than a less frequent high dose. In a previous study, JF-10-81 did not affect human prostate cancer PC-3 tumor growth using a less frequent low dose protocol (0.73 mg/kg, twice a week) [9]. However, current results indicate that JF-10-81 obviously inhibited PC-3 tumor growth at the more frequent higher dose (3.5 mg/kg, 3 times a week), resulting in an almost 60% decrease in volume (Fig. 5).
Taking these observations into account and the expectation that the targeted conjugate would have an improved therapeutic/toxicity ratio, an approach using a continuous drug-releasing pellet of 60 days was implemented. The results show that this new continuous drug-releasing approach results in signifi cantly improved tumor responses and with no evidence of animal body weight losses. JF-10-81 resulted in an over 85% reduction in tumors of neuroblastoma IMR32 and leukemia MOLT-4 with solid tumor regression and even disappearance of some CFPAC-1 tumors occurring. As for SSTRoverexpressing, but CPT-insensitive BON tumors, JF-10-81 exhibited much more potent inhibitory effects with over double the inhibitory rate seen with the CPT formulation perhaps by increasing the CPT concentration within the vascularized solid tumor or possibly inside the tumor cells themselves (data not shown). And previously JF-10-81 also displayed anti-angiogenic activity in vitro [25], and now additionally showed its ability to inhibit the metastasis of the highly invasive PC-3 tumors (Fig. 5). Therefore, this conjugate JF-10-81 with optimal drug-releasing approach could be a new potential chemotherapeutics. Reportedly, the cultured tumor cells IMR32 and CFPAC-1 highly express some SSTRs [33,39,40], with BON cells natively expressing high levels of SSTR1, SSTR2 and SSTR5 [41,42]. Our results are similar to the previous reports. Abundant SSTR2 mRNAs were detected in all 5 cultured tumor cells and SSTR2 proteins were detected as well. However, except BON and IMR32 cells, we observed little or very weak binding affi nities of total 5 SSTRs in CFPAC-1, MOLT-4, PC-3 and many other tumor cells. Somatostatin binding has been demonstrated in tumor tissues taken from various primary tumors including neuroblastoma, pancreatic cancer, carcinoid and leukemia and prostate cancer [14]. Part of this problem may lie in the over passaging of commercially available cell lines and we have observed that native tumor cells and even receptor-transfected cells with initially high binding affi nity can gradually lose their receptor density and binding specifi city and rise non-specifi c binding after multiple passages. There is the possibility that cultured cells lose their natural in vivo microenvironments that can stimulate expression of SSTRs and their functional appearance on cell surfaces. This and additional interactions with other cells or growth factors could be part of several possible explanations for why JF-10-81 had greater in vivo but less in vitro effects than naked CPT. Additionally, in animals the conjugate would have the ability to target and kill SSTR-positive tumor blood vessels thus providing an additional mechanism for inhibition of tumor growth [10].
In conclusion, this CPT-SST conjugate has been shown to effectively inhibit angiogenesis and the progression of various SSTR-positive tumors, with low toxicity. Apparently, through targeting receptorspecifi c tumor sites, this kind of novel cytotoxic peptide conjugates can increase cancer cell selectivity and cytotoxicity and reduce toxic side effects relative to the traditional chemotherapy.