Microencapsulated octreotide pamoate in advanced gastrointestinal and pancreatic cancer : a phase I study

Fourteen patients suffering from advanced colorectal (n = 7), pancreatic (n = 4) or gastric (n = 3) carcinomas received treatment with microencapsulated octreotide pamoate 90 mg i.m. every 4 weeks (n = 4), 160 mg i.m. every 4 weeks (n = 4) or 160 mg i.m. every 2 weeks (n = 6). Two patients had stable disease, one for 4 and one for 6 months. Plasma insulin-like growth factor (IGF)-I decreased by 49-53%, IGF-II by 27-37% and total IGF-binding protein (IGFBP)-3 by 16-19%, whereas IGFBP-1 increased by 35-55%. Insulin and C-peptide levels decreased by 29-38% and 41-46% respectively. A non-significant decrease in urinary GH secretion and an increase in the ratio of fragmented to intact IGFBP-3 as well as IGFBP-3 protease activity was seen. The increase in IGFBP-3 fragmentation correlated negatively with alterations in IGF-I and IGF-II (P < 0.05). We conclude that microencapsulated octreotide administered in doses up to 160 mg every 2 weeks is well tolerated and has pronounced effects on several components of the IGF system in plasma. In addition, changes in IGFBP-3 protease activity because of cancer may contribute to alterations in IGF-I and -II, indicating the importance of measuring this parameter in addition to IGFs and IGFBPs when evaluating alterations in IGF-I. ImagesFigure 2

are needed to improve therapy. Somatostatin analogues are found to be useful in palliative treatment of certain endocrine gastrointestinal tumours (Schally. 1988). In vitro and animal studies have also shown somatostatin analogues to inhibit the growth of nonendocrine pancreatic and gastrointestinal tumours (Szepeshazi et al, 1991;Dy et al. 1992: Qin et al. 1992: Watson et al. 1992. Although grow%th inhibition by somatostatin analogues may be a direct effect mediated by specific somatostatin receptors on tumour cells. such effects could also be achieved through inhibition of growth factors and growth stimulatory hormones. Somatostatin analogues inhibit secretion of gastrointestinal hormones with possible mitogenic effects (Adrian et al. 1981 ) and suppress synthesis of the insulin-like growth factor (IGF)-I (Pollak et al. 1989: Figg et al. 1995: Leo et al. 1995 by suppressing growth hormone (GH) secretion (del Pozo et al. 1986). the main trophic factor for IGF-I synthesis (Schwander et al. 1983). IGF-I and -H are important mitogens to many human cancer cell lines in vitro and altered bioavailability of these growth factors may potentially influence tumour growth (Macaulay, 1992).
Clinical studies evaluating the anti-tumour effect of somatostatin analogues in non-endocrine abdominal cancers have reported conflicting results (Savage et al. 1987: Klijn et al. 1990: Friess et al. 1993: Smith et al. 1994. Cascinu et al. 1995. A major obstacle to the development of somatostatin analogues in clinical use has been their short half-lives. demanding three daily injections (Kutz et al. 1986). In this study we evaluate the tolerability. clinical effects, pharmacokinetics and effects on plasma IGF-I and -II of a new microencapsulated formulation of octreotide pamoate in patients suffering from gastrointestinal and pancreatic carcinomas. Plasma IGFs are bound to specific IGF-binding proteins (IGFBPs) acting as carriers and also as modulators of IGF-I bioactivity in the tissues (Jones and Clemmons. 1995). Most IGF-I and -H1 circulate in a 1 50-kDa complex together with an acid-labile subunit (ALS) and IGFBP-3. and availabilit) of these growth factors to the tissues may be affected by proteases acting on IGFBP-3 as substrate . Thus. apart from determining plasma IGF-I and -HI together with IGFBP-1 and -3 with radioimmunoassay (RIA). we measured IGFBPs using Western ligand blotting and the functional status of IGFBP-3 using immunoblotting. In addition. we also measured IGFBP-3 protease activity in the plasma samples.

PATIENTS AND METHODS Patients
Fourteen patients (eight men and six A-omen) with progressive locally advanced and/or metastatic adenocarcinomas of abdominal origin were studied (Table 1). Any other svstemic anti-cancer therapy had been terminated at least 4 weeks before inclusion in the study. All patients had a performance status (WHO) less than 2 and an expected survival of more than 3 months. None of the patients had any significant medical or surgical disorder apart Part of this s-ork sA-as presented at the 21 St ESO1 congress in %ienna 2-5 Nos ember 1996.
Phase I study of octrotde pamoate 15

Treatment schedule
Patients were assigned sequentially to one of three cohorts. The first cohort received octreotide pamoate (OncoLAR) 90 mg i.m. every 4 weeks, whereas cohorts 2 and 3 received octreotide 160 mg i.m. every 4 weeks and every 2 weeks respectively. Four to six patients were enrolled in each cohort, and the firquency of dose-limiting toxicities (grade 3) should be < 1 in four subjects or < 2 in six subjects to allow patients to be enrolled in the next cohort. Thus, four patients were treated in cohorts 1 and 2, whereas six patients were trated in cohort 3. Physical examination, evaluation of adverse effects and standard laboratory evaluations (haematology, chemistry and urine analysis) were performed weekly until first response evaluation at day 57. Response evaluation was carried out radiologically according to the UICC criteria.
Patients showing either stable disease or an objective response were offered continuous treatment with octreotide pamoate in the same dose as they had rceived during the first 8 weeks of the study with weekly follow-up for toxicity. Patients in cohorts 1 and 2 not continuing maintenance therapy had follow-up visits every 4 weeks, whereas patients in cohort 3 had extended weekly examinations up to day 71 after commencing treatment and later every 4 weeks.

Bklod and unne sampling
Fasting blood samples for evaluation of the IGF system, insulin and C-peptide were obtaied in heparinized vials on the morning of commencing treatment and subsequently on days 15, 29, 57, 71, 99, 127, 155 and 169 during the reatment period. Plasma was separated by centrifugation and stored at -20°C until analysis.
Overnight (12 h) urine was collected on the same days for estimation of GH secretion. Serum samples for measurement of octroide levels were obtained weekly during the study period. At the day of drug administration additional samples were drawn before and 1, 2 and 3 h after injection.

Assays
Plasma levels of IGF-I (Holly et al, 1988) and IGF-lI (Davies et al, 1991) were measured using RIA after acid-acetone extraction. IGFBP-3 (Cwyfan-Hughes et al, 1993) and IGFBP-1 (Holly et al, 1988) and octreotide (del Pozo et al, 1986) were directly measured using RIA. Plasma insulin and C-peptide were measured using RIA kits purchased from Diagnostic Products Corporation (Los Angeles, CA, USA). Urinary GH was measured by a sensitive RIA kit obtained from BioMerieux (France) according to the manufacturer's instructions. The IGFBP profile in the plasma was analysed by Western ligand blotting (WLB) using a modified version  of the techniquw orginally descnrbed by Hossenlopp (Hossenlopp et aL 1986). IGFBPs were visualized by autoradiography and quantified using a densitometric scanner (Pharmacia LKB, Uppsala, Sweden).
The IGFBP pattern was compared with the profile of a noml plasma pool (NP), and samples from each patient were analysed in the same run for comparson.
After WLB the membranes were washed and blocked four tmes in 10 nm Tris-HCl (pH 7.4) containing 5% milk and 0.2% Tween 20. The membranes were then probed overnight with a polyclonal specific antiserum against IGFBP-3 purchased from Diagnostic Systems Laboratories (Webster, TX, USA) at a final dilution of 1: 10000. The membranes were then developed using enhanced chemiluminescent reagents supplied by Amersham (Aylesbury, UK) according to the manufacturer's instructions and the films were subjected to densitometric scanniing.
The IGFBP-3 proteolytic activity in plasma was examined using a modified version  of the technique described by Lamson et al (1991).

Statistics
Cohort 1 In pevious studies we found plasma levels of IGF-I and -11, and 50 IGFBP-1 to be well fitted to a log normal distribution whereas IGFBP-3 was found to be nomally distributed (L0nning et al, 1995; Helle et al, 1996). Thus, parameters are given as their geometrical 40mean value with 95% confidence intervals of the mean, with the E exception of IGFBP-3 and plasma octreoide levels when the arithmetic mean values are given. Correlations between parameters were X 30tested for using the Spearman rank correlation test.  . rlevel (98.9 ng ml-') at day 13 with a subsequent gradual decline. peptide measured using RIA before the first injection on day l and Day of treatmet on tratment levels expressed as percentage of pretreatment values are shown in Table 2. No differences between the cohorts or diagnostic subgroups were observed for any parameter. Accordingly, FIgue 1 Pma evels of rareide dunng tde ve beFree coort all data were pooled for statistical analysis. The number of patients gven as mea lvels wWit range. Only octeohde beel before qedion are shown at days of drug abstraton. Vkues obtained 1, 2 and 3 h r available to follow-up beyond day 57 was too small to permit any irectlon are not ickuded statistical analysis of these data Phase I study of octreotide pamoate 17 Table 2 Plasma levels of IGF-1, IGF-I1, IGFBP-1, IGFBP-3, insulin, C-peptide and urinary GH excretion at day 1 (before the first injection) and levels at different time intervals during treatment with octeofide pamoate given as a percentage of pretreatment levels. Values other than IGFBP-3 RIA are given as geometrical FKgure 2 Western ligand blots (above) with corresponding immunoblts for IGFBP-3 (below) in samples from two patients before and dunng treatmnent with octreotde pamoate. The 42to 44-kD band on the ligand blot corresponds to the two gycsylation forms of intact IGFBP-3, the 364kDa and 244kDa bands correspond to IGFBP-2 and IGFBP-4 respectively. The nature of the 30to 324kDa band was not established. Patient 1 had a rapid progressive disease and an increase in the ratio of fragmented to intac IGFBP-3, whereas patient 2 had a slower disease progression and small aiterabons in protease actvity. NP, normal plasma Plasma IGF-l levels decreased by 49-54%. whereas IGF-H decreased by 27-37% at various time intervals on treatment. We also observed a moderate decrease (16-19%) in immunoreactive IGFBP-3. IGFBP-1 levels increased by 35-55%. whereas fasting levels of insulin and C-peptide decreased by 29-38% and 41-46% respectively. A negative correlation between alterations in plasma C-peptide or insulin and IGFBP-1 was observed at most time intervals, but except for the correlation between C-peptide and IGFBP-1 at day 15 (P < 0.05) none of these correlations was of statistical significance.

10-
IGFBP-3 was evaluated by Western ligand blots and immunoblots in addition to RIA (Table 2). Densitometric scanning of ligand blots revealed a mean decrease in intact IGFBP-3 between 26% and 49%. This decrease was somewhat larger than that observed in total IGFBP-3 evaluated using RIA. Although we observed only a minor overall increase in the ratio of fragmented to intact IGFBP-3 and in IGFBP-3 protease activity, several patients had a substantially higher ratio of fragmented to intact IGFBP-3 in the samples obtained at the time when disease progression was recorded (Figure 2). A positive correlation between the decrease in IGF-I and -H plasma levels and increase in fragmented to intact IGFBP-3 was found at all time interval. but it was of statistical significance only on day 57 (P < 0.05). Densitometric scanning of low-molecular-weight IGFBPs on the British Joumal of Cancer (1998) 78(1) One patient received only one injection of o patient all parameters retuned to baseline levels at DISCUSSION Although previous somatostatin analogue formul found to be generally well tolerated, two to three i were required because of their short half-lives (I Plasma steady-state octreotide concentrations of C are reported to cause a 70-80% inhibition of ar} growth hormone secretion in healthy humans 1992). However, much higher plasma levels of 500 ng ml-') have been reported to be necessary vivo growth inhibition of colon cancer xenogra (Dy et al, 1992). The lowest value of octreot individual patients during treatment in our study whereas the mean levels for each cohort were cor at different time intervals. Moderate variations and low peak concentrations suggest adequate d i.m. administered octreotide pamoate. Although t all patients achieve plasma octreotide levels suffic GH suppression in healthy individuals, the coi different in patients suffering from advance patients have been reported to have eleva (Emermann et al, 1984;Klijn et al, 1990). The stel mean plasma levels of octreotide in cohorts 2 1 weeks) and 3 (160 mg every 2 weeks) until day steady state was not reached before terminating tion in the majority of the patients. Owing tc absorption from the depot formnulation, and the drug availability after OncoLAR administratioi common pharmacokinetic variables such as te clearance rate and volume of distribution could n in this study.
We found octreode pamoate administered in doses up to 160 mg every second week to be well tolerated in patients with advanced cancer. The overall incidence of side-effects was low in all cohorts, and no grade 3 side-effects were observed. Treatment with octreotide pamoate had pronounced effects on the IGF system. The 50% decrease in plasma levels of IGF-I is in accordance with recent reports on somatuline (Figg et al, 1995) and lanreotide (di Leo et al, 1995) as well as a previous study with octreotide (Pollak et al, 1989). Mean baseline levels of IGF-1 (65 ng ml-') and IGF-H (275 ngml-') were much lower than the normal range of these peptides (IGF-1; 100-494 ng ml-' and IGF-IGF-I I1; 462-1042 ng ml-')used as reference by others (Lawrence et al, *@-- IGF-lI 1997). This may be explained partly by advanced age as well as *U-BP-3 weight loss in our patients, both known to decrease plasma IGF-I levels (Sara and Hall, 1990). It cannot be excluded that some of the futher decrease in IGF-I and IGF-HI observed beyond day 15 may 80 100 be due to disease-related factors, but only a minority of patients experienced major weight loss or deteriorating liver function.
ing R napaent Te drop in plasma IGF-I may bedueto inhibitionof GH secremtm Pne The tion by octreotide causing decreased hepatic synthesis of this s in the IGFBP-3 growth factor (Schwander et al, 1983), but other mechanisms may also operate (Serri etal, 1992). The observed 27-37% decrease in IGF-ll levels was unexpected, as previous studies with the somatostatin analogue somatuline (Figg et al, 1995) as well as other om a progressive hormonal therapies influencing IGF-I levels (Frost et al, 1996; % to 40% in the Reed et al, 1992) were found to have no effects on IGF-H. We lwn).
speculate that some of the decrease in IGF-ll levels may be ctro ide. In this secondary to a GH-dependent decrease in IGFBP-3 in our patients. d ay 84 (Figure 3). GH is the most important regulator for synthesis of ALS (Dai et al, 1994), which, together with IGFBP-3, is necessary for formation of the 150-kDa ternary complex. A decrease in ALS (not evaluated in our study) affecting formation of the terary complex may l ations have been subsequently reduce the amount of IGFBP-3 as well as available i njections per day binding sites for IGFs in this complex. The possible explanation K utz et al. 1986). may he that low-molecular-weight complexes consisting of only .27-0.55 ng ml-' IGF-I or -i and an IGF-binding protein (including IGFBP-3) have paine-stimulated a shorter half-life than the ternary complex (Guler et al, 1989). The (Marbach et al, decrease in IGFBP-3 is probably not secondary to the fall in IGF-1, octreotide (about as a previous study has shown that although administration of GH for significant in increases plasma IGFBP-3 levels it was decreased by administrafts in nude mice tion of IGF-I (Kupfer et al, 1992). ide measured in Although most of the observed alterations in IGF-1, -II and was 3.4 ng ml-', IGFBP-3 may be explained by effects of uratment with octreotide, isiderably higher an increase in IGFBP-3 protease activity related to disease in plasma levels progression may also influence these parameters. Although total lepot function of IGFBP-3 determined using RIA decreased by 15-20% only, this suggests that densitometric scanning of IGFBP-3 on WLB revealed a larger ient for effective suppression of intact IGFBP-3 (26-49%). Most IGFBP-3 RIAs nditions may be detect both intact IGFBP-3 as well as fragments, and the observed cancer, as many increase in IGFBP-3 protease activity is not reflected by the RIA ited GH levels results.
Discrepancies between RIA and have also been pwise increase in observed by others (Gargosky et al, 1992), indicating the impor-(160 mg every 4 tance of evaluating IGFBP-3 also by WLB and immunoblots. The 57 indicates that decrease in IGF-I and -H was also positively correlated with an drug administraincrease in IGFBP-3 fragmentation at day 57, indicating that alterthe continuous ations in IGFBP-3 protease activity may also influence plasma fact that plasma levels of IGF-I and -H. An increase in IGFBP-3 protease activity n is not known, been reported to increase plasma clearance in IGF-I in rats .rminal half-life, (Davenport et al, 1990 IGFs to the tissues. The IGFBP-3 protease activity has been reprted to be high in GH-deficient patients and low in acmegalic patients, consistent with an inverse relationship with IGFlevels (Lassarre et al, 1994). Our finding of an increase in fasting plasma levels of IGFBP-l is in accordance with the findings of others (Ezzat et al, 1992;Wolthers et al, 1994). Plasma IGFBP-1 is inversely correlated with insulin levels in healthy subjects, and insulin is known to be one of the most important regulators of plasma IGF'BP-I (Holly et al, 1988). Data from some studies indicate an insuln-independent regulation of IGFBP-l by somatostatin analogues (Ezzat et al, 1992;Wolthers et al, 1994). However, other investigators have reported an inverse correlation between insulin and IGFBP-l also during reatment with octreoide (Fredstorp et al, 1994), and hyperinsulinaemia was found to abolish somatostatin-stimulated IGFBP-1 release (0rskov et al, 1994). In this study, fasting plasma levels of both insulin and C-peptide were significantly decreased and correlated negatively with alterations in plasma levels of IGFBP-1. Thus, our data support a role of insulin in the regulation of IGFBP-1 during treatment with octreotide. It is difficult to assess the influence of an increase in IGFBP-1 on the biological actions of IGF-I as this binding protein may inhibit but also enhance IGF-I effects in vivo, depending on the phosphorylation status of this binding protein (Jyung et al, 1994).
Densitometric scanning of WLB revealed an increase in the 36-kDa band corresponding to IGFBP-2. In a previous study, IGFBP-2 was found to be increased in many cancer patients (Kanety et al, 1993), and it is possible that the increase in this binding protein observed during tratment may be associated with disease progression rather than any influence of drug treatment. Somatostatin analogues are known to inhibit GH secretion (del Pozo et aL 1986). Surprisingly, we did not observe a consistent suppression of urinary GH in our patients. Urinary 12-h GH secretion was measured using a sensitive RIA, and previous studies have shown a good correlation between urinary GH secretion and plasma GH profile (Girard and Fischer-Wasels, 1990;Hourd and Edvards, 1989). However, only a small amount (< 0.01%) of plasma GH is normally excreted in the urine (Baumann and Abramson, 1983), and whether this fraction may change in patients with advanced cancer is not known. We observed particularly high values in the last urinary samples obtained in patients with rapid progressive disease. Thus, the validity of urinary GH measurements in patients suffering from advanced cancer may be questionable. But it is also possible that no clinically significant suppression of plasma GH is obtained by treatment with octreotide in these patients as has been observed by others (Klijn et al, 1990).
In conclusion, octreotide pamoate in doses up to 160 mg every second week provides high plasma drug levels and is well tolerated in patients with advanced cancer. All doses administered were found to significantly suppress plasma levels of IGF-I, IGF-II, IGFBP-3, insulin and C-peptide, and to increase plasma levels of IGFBP-1. Whether suppression of plasma IGF-I and -II may be of importance regarding tumour growth is not yet clear. The durable endocrine effects and clinical tolerance suggest that octreide administered as its pamoate depot formulation should be evaluated in further trials in cancer patients.

ACKNOWLEDGMENTS
The technical assistance of Mr D Ekse is highly appreciatd. We are grateful to Celtrix P"harmaceuticals for the generous provision of the recombinant IGFBP-3 used in the assays. We would also like to thank Dr Christophe Gerbeau (Pharnacokinetic Unit, Laboratoire Sandoz, Rueil Malmaisone, France) for performing the octreotide assays. This work was supported by grants from the Norwegian Cancer Society.