Phase I Trial of Orally Administered Pentosan Polysulfate in Patients with Advanced Cancer1 in Vitroand

Vol. 3, 2347-2354. Decel?lber /997 Clinical Cancer Research 2347

Phase I Trial of Orally Administered Pentosan Polysulfate in Patients with Advanced Cancer1

John L. Marshall,2 Anton Welistein, James Rae, dose was increased, but their seventies were similar at all Robert J. DeLap, Kim Phipps, John Hanfelt, dose levels. There were no objective responses, although three patients had prolonged stabilization of previously pro- Manasses K. Yunmbam, Jim X. Sun, gressing disease. Pharmacokinetic analysis suggested Kenneth L. Duchin, and Michael J. Hawkins marked accumulation of PPS upon chronic administration. Lombardi Cancer Research Center, Georgetown University Hospital. Serum and urine bFGF levels failed to show a consistent, Washington. DC 2X)7 Ii. L. M.. A. W.. J. R.. K. P.. J. H.. M. J. H.J: Baker Norton Pharmaceuticals. Inc.. Miami. Florida 33137 Ii. X. S.. interpretable pattern; however the data suggested an in- K. L. D.j: Division of Oncology Drug Products. Center ftr Drug verse relationship between PPS and bFGF levels in s’ivo. A Evaluation and Research. Food and Drug Administration. Rockville. MTD could not be determined using the daily t.i.d. dosing Maryland 20854 lR. J. Dl: and NIH. National Cancer Institute. schedule due to the development of grade 3/4 GI toxicity Bethesda. Maryland 20892 [M. K. Y.] (proctitis) at all dose levels studied. PPS, administered p.o. at doses of 400 mg/m2 t.i.d., did not cause significant sys- ABSTRACT temic toxicity, but most patients developed moderate-to- Tumor angiogenesis is critically important to tumor severe GI toxicity within 1-2 months. The cause of the GI growth and metastasis. We have shown that pentosan poly- toxicity was unclear, but it was readily reversible upon sulfate (PPS) is an effective inhibitor of heparin-binding cessation of the agent. The suggestion of an inverse relation- growth factors in vitroand can effectively inhibit the estab- ship between PPS and bFGF supports further study of PPS lishment and growth of tumors in nude mice. Following as an antiangiogenic agent. The tested doses and schedule completion of our Phase I trial of s.c. administered PPS, we cannot be recommended for further study. Subsequent mu- performed a Phase I trial of p.o. administered PPS in pa- rine experiments showed PPS to be more effective as an tients with advanced cancer to determine the maximum anticancer agent when it is given intermittently. We propose tolerated dose (MTD) and toxicity profile and to search for a study of PPS given on a weekly schedule in further clinical any evidence for biological activity in vivo. Patients diag- trials. nosed with advanced, incurable malignancies who met standard Phase I criteria and who did not have a history of INTRODUCTION bleeding complications were enrolled, in cohorts of three, to As reviewed by Folkman ( I ). it is becoming increasingly receive PPS p.o. t.i.d., at planned doses of 180, 270, 400, 600, evident that angiogenesis (new blood vessel formation is ot and 800 mg/rn2. Patients were monitored at least every 2 critical importance in the growth and metastatic spread of solid weeks with physical exams and weekly with hematological, tunors. Tumor growth in animal model systetils is limited in the chemistry, stool hemoccult, and coagulation blood studies, absence of vascularization, and inhibitors of angiogenesis that and serum and urine samples for PPS and basic fibroblastic are not cytotoxic or cytostatic in vitro can inhibit tumor growth growth factor (bFGF) levels were also taken. The PPS dose in l’il() (2-6). Furthermore. neovascularization may be a critical was escalated in an attempt to reach the MTD. Eight addi- determinant of the metastatic potential of a tumor. Tumor cells tional patients were enrolled at the highest dose to further may not gain entry into the systemic circulation in significant characterize the toxicity profile and biological in vivo effects numbers until tumor neovascularization has occurred, and of PPS. A total of 21 patients were enrolled in the three larger. more vascular tumors may shed more tumor cells into the cohorts of doses 180 (ii = 4), 270 (ii = 3), and 400 (ii = 14) bloodstream (7-9). Clinically. cancers often demonstrate a pre- mg/rn2. The most severe toxicities seen were grade 3 proctitis vascular or in situ phase and a more advanced vascular phase and grade 4 diarrhea; however, 20 of the 21 patients had (10-12). The appearance of’ angiogenic activity in a histologi- evidence of grade 1 or 2 gastrointestinal (GI) bleeding. These cally benign or i,i situ malignant lesion may be an important toxicities became evident at a much earlier time point as the indicator that neovascularization and transition to a more malignant. invasive cancer will occur (13-17). Clinical studies have shown a positive relationship between the degree of tumor vascularization and metastatic spread (18, 19). Clinically. inter- Received 1/28/97: revised 9/1 1/97: accepted 9/29/97. ference with tumor angiogenesis could serve to delay or arrest The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked disease progression in cancer patients. Because active angiogen- advertise,ne,,t in accordance with 18 U.S.C. Section 1734 solely to esis in normal tissues of adults occurs only in certain physio-

indicate this fact. logical situations (e.g. . wound healing or the response of the I This work was supported by NIH Grant 5 VOl CA62500. uterine endornetriurn to hormonal stimulation). antiangiogenic 2 To whom requests for reprints should be addressed. at Lombardi therapy would be expected to have limited adverse effects on Cancer Research Center. Georgetown University Hospital. 3800 Reser- voir Road, NW.. Washington. DC 200()7. E-mail: marshalj@gunet. normal tissues. georgetown.edu. Angiogenesis is physiologically controlled by growth fac-

tors and growth factor modulators. HBGFs3 are particularly plasma and urine levels were higher than they were after the first potent in stimulating motility and proliferation of endothelial dose (Baker Norton Pharmaceuticals, Inc.). cells and inducing angiogenesis in several laboratory model We previously reported a Phase I study of PPS, adminis- systems (20-28). The HBGF family includes several known tered q.i.d. by s.c. injection, in patients with advanced cancer. members, which are now denoted HBGF-1-HBGF-7 (29, 30). Nineteen patients were enrolled in cohorts using escalating In addition to their effects in stimulating endothelial cells and doses from 60 to 120 mg/m2 q6h. The DLT was reversible angiogenesis, HBGFs have demonstrated mitogenic and differ- thrombocytopenia; also noted were prolonged activated partial entiation-inducing effects on several other types of normal cells thromboplastin time, hematuria, transient increases in transami- (31, 32). nases, bleeding from uterine fibroids, and rash. The recom- Studies have shown that malignant tumor cells commonly mended Phase II dose was 90 mg/m2 q6h. Although a few secrete substantial quantities of angiogenic growth factors, in- patients may have experienced stabilization of their previously cluding HBGFs. We have shown that PPS (also called pentosan) progressive malignancies, most of the patients with advanced is an effective inhibitor of HBGF-mediated cellular proliferation cancer enrolled in this study appeared not to benefit from PPS

in vitro and can effectively inhibit establishment and growth of treatment (41 ). Plasma samples from patients were found to tumors in athymic mice that have been injected with tumori- have prolonged, high levels of circulating anti-HBGF activity, genic human cancer cells (33, 34). However, in other animal potentially sufficient to inhibit angiogenesis in microscopic tu- studies, PPS had little or no therapeutic effect on the progression mor deposits in vivo, with only modest effects on coagulation of established tumors. This is presumably because established parameters (42-44). tumors have already formed a blood supply and/or because On the basis of our experience to date, PPS may be most large, established tumors may elaborate very high levels of useful as an agent to suppress progression of cancer in patients growth factors, which cannot be effectively blocked by clini- who have a low tumor burden. For optimal results, chronic administration of PPS (possibly several times daily) for months cally tolerable doses of PPS. or years may be necessary. To date, most clinical pharmacolog- Because of its activity in retarding the blood coagulation ical studies of PPS have evaluated the s.c. route of injection. cascade (35), PPS is marketed outside the United States for use Although administration of PPS by repeated s.c. injection or by as a heparin-like anticoagulant in doses from 50 to 100 mg s.c. continuous infusion may be technically feasible, p.o. adminis- t.i.d., and it results in not only the expected toxicity of bleeding tration (if clinically safe and efficacious) would clearly be complications but also, rarely, thrombocytopenia, elevated preferable for long-term therapy. Published data suggest that transaminases, and thromboembolic reactions (36). Other din- p.o. administration of PPS may indeed be feasible, at least for ical trials have explored the role of PPS in the treatment of noncancer clinical indications. Here, we systematically evalu- chronic interstitial cystitis and chronic radiation-induced prod- ated the clinical, pharmacological, and in vivo biological effects titis (37). In prior studies of the treatment of interstitial cystitis, and the safety of escalating doses of p.o. PPS. PPS was supplied PPS has been administered p.o. at 100-300 mg t.i.d., for 3 to by Baker Norton Pharmaceuticals and was administered under > 18 months (38). As demonstrated by double-blind trials in an Investigational New Drug application held by the Lombardi interstitial cystitis, PPS was effective in alleviating the symp- Cancer Center. toms of interstitial cystitis and was very well tolerated, with few adverse effects reported (38, 39). In particular, clinically significant anticoagulation has not been reported in these studies. PATIENTS AND METHODS Although p.o. PPS appears to be clinically effective in the From August 1993 to March 1995, 21 patients were en- treatment of interstitial cystitis, there is very little published rolled in our Phase I trial of p.o. PPS. The initial phase of the information regarding the quantitative p.o. bioavailability of trial involved a standard dose escalation in cohorts of at least PPS. In healthy volunteers, approximately 3% of a p.o. dose of three patients. At the highest dose level achieved, additional tritiated PPS was excreted unchanged in the urine, compared to patients were enrolled to better characterize the toxicities of this renal excretion of parent drug of 8-10% following an iv. dose dose and to gather additional biological and pharmacokinetic of PPS (40). This would imply a p.o. bioavailability in the range data at a fixed dose. of 30% or more, clearly compatible with clinical use of the p.o. Patient Eligibility. All patients in this trial required a route of administration. In contrast, Danielson et a!. (40) re- pathologically confirmed diagnosis of metastatic malignancy; ported that the ratio of 24 h urinary excretion of PPS, assayed by no known history of other modes of therapy that would produce competitive binding, was about 0.007 after p.o. (400 mg) and a chance of cure or significant palliation; performance status of iv. (40 mg) administration in eight healthy volunteers of PPS on 0 or 1 ; age of 18 years; availability for frequent follow-up; a q8h regimen, suggesting that p.o. bioavailability was <1% weight loss of 10% in the last 28 days preceding the trial; after a single dose. With multiple dosing in healthy subjects, recovery from any prior toxicities of chemotherapy, radiation, and/or surgery; no clinical evidence of active brain metastasis; expected survival of 8 weeks without therapy; no serious intercurrent illness that would interfere with the treatment pro- gram; negative 3-human chorionic gonadotropin and effective 3 The abbreviations used are: HBGF, heparmn-binding growth factor; means of birth control (women of child-bearing potential); and PPS, pentosan polysulfate; DLT, dose-limiting toxicity; GI, gastrointestinal; MID, maximum tolerated dose; NSB, nonspecific binding; bFGF, ability to understand and give informed consent. This study was basic fibroblast growth factor. performed after approval by the Georgetown University Insti-

tutional Review Board. Physiological parameter requirements dose, including the 6 patients who were enrolled at 400 mg/rn2 included WBC counts of 3,000/mm3, absolute neutrophil during the initial phase of the trial. counts of 1 ,500/mm3, platelet counts of 100,000/mm3, cre- PPS Assay Methods. A validated RIA was used to quan- atinine of 1 .5 mg/dl or creatinine clearance of 60 ml/min. titate PPS in plasma. Antibodies were raised against PPS by and total bilirubin of 1 .5 mg/dl. In addition, patients could not coupling PPS to methylated BSA. This complex was hornoge- have any history of cerebral vascular hemorrhage or thrombosis, nized with Freund’s adjuvant and injected into New Zealand recent GI or urological hemorrhage, bleeding or platelet disor- white rabbits (1 mg/rabbit). The methodologies used to prepare ders, congestive heart failure, pericardial effusion. or uncon- and quantitate the antibodies were described by Callahan et al. trolled hypertension or require heparin, warfarin, nonsteroidal (45). The plasma or urine sample (75 p.1), tyrosine-PPS labeled anti-inflammatory agents, or other antiplatelet agents. Although with 1251 and diluted antiserum were mixed together. After a most patients had measurable or evaluable disease, this was not 2-h incubation at 37#{176}Cto establish the competitive-binding a requirement for entry onto the study. Because of the nature of equilibrium, a second antibody directed against rabbit IgG was this agent and the proposed mechanism of action, patients with added. After 30 mm, the precipitated IgG was separated by stage IV malignancy but without evidence of macroscopic dis- centrifugation. The supernatants were then removed by con- ease (i.e. , stage IV, no evidence of disease) were eligible for trolled aspirations. The radioactive iodine in the pellet was then treatment and evaluable for toxicity and for disease progression determined using a commercial counting system. Sample con- but not for response. centrations were calculated by reference to a fitted calibration Treatment and Dosage Escalations. Thirteen patients curve using quality control standards. Each known or unknown were enrolled into the initial dose escalation phase of the trial in sample was assayed in duplicate. For each assay, standards of cohorts of at least three patients and were given PPS, beginning PPS at 2, 4, 6, 10, 20, 40, and 60 ng/ml in human plasma or with a dosage of 180 mg/m2 t.i.d., escalating to the MTD of 400 urine were used. Process control samples of PPS at concentra- mg/m2 t.i.d. The initial dose of 180 mg/m2 was chosen to tions of 8.50 and 500 ng/ml were also used. NSB was deter- approximate 3 times the dose of PPS that was given for pro- mined by using normal rabbit serum rather than specific anti- longed durations in interstitial cystitis trials (39). Dose escala- body. A validation curve consisted of a set of standards, tions occurred if PPS was well tolerated for 4 weeks in all three controls, and NSB samples, all in duplicate. A validation set patients entered at the previous dose level. The agent was consisted of three such curves processed on the same working administered p.o., every 8 h, on days 1-56. Patients were eval- day. As part of the assay validation procedure, three sets of three uated weekly for toxicity by physical examination, complete curves each were run on 3 separate days. The concentrations of blood count and differential, prothrombin time/partial thrombo- PPS were computed with the following parameters: number of plastin time, stool guaics for occult blood, and chemistry panel. counts bound, number of counts bound in the absence of added Pharmacokinetic samples were obtained on days 1 and 15. sample or standard, and NSB. The data were fitted to a linear Tumor assessments were performed just prior to the day 56 regression to determine the concentration of PPS in the sample. clinical evaluation. Responses were defined as either progres- The mean calculated concentrations for each of seven PPS sive disease (25% increase in bidimensional measurements) or standards (2-60 ng/rnl) ranged from - 8.3 to 1 1 .9% of the target stable disease (neither a response nor progression, lasting at value. Results of freeze-thaw of cycling samples were also least 3 months). Obviously, only stable disease was possible for examined. Freezing and rethawing of the samples, up to 3 times, patients with disease classified as stage 4 with no evidence of provided concentrations of PPS that were 94-105% of the target disease. In the absence of progressive disease, patients with concentration. either a response or stable disease could continue the therapy. bFGF Analysis Methods. Urine and serum samples with monthly reassessments. were obtained at the specified time points and analyzed using an MTD. At least three patients were enrolled into each ELISA (R&D Systems, Inc.). Concentrations from samples cohort. If one of the three patients experienced a DLT during the were determined by extrapolating from standard curves. initial 4 weeks of therapy, three more patients were enrolled at Statistical Methods. To determine whether there was a that dose level. If no further DLTs were seen, the subsequent correlation between bFGF and PPS levels, measured at the same three patients would be enrolled in the next higher dose level. If time on day 1 of treatment, we fitted a weighted linear regres- two of three patients on a given dose level experienced a DLT sion model. Weights were selected to maximize the precision of within the first 4 weeks of therapy, the next three patients would the estimate; i.e. , weights were proportional to the inverse of the be enrolled into the next lower dose level. The MTD was measurement variance of the bFGF assay. The General Estimat- defined as the highest dose that resulted in no more than one ing Equations method was used to adjust for repeated measures patient of six experiencing a DLT. The planned dose levels were within a patient (46). 180, 270, 400. 600, and 800 mg/rn2. with subsequent escalations in 25% increments. Accrual of Additional Patients. Because of the devel- RESULTS opment of readily reversible DLT that was universal but oc- A total of 21 patients were enrolled into this Phase I trial of cuned sooner at the higher dose levels, eight patients were then P.O. PPS. Patient characteristics are listed in Table I . The dose enrolled at the highest dose (400 mg/rn2 t.i.d.) to better charac- levels and the numbers of patients treated at each level are listed terize the toxicity of the dose level and the biological activity of in Table 2. The details of each dose level are expanded below. PPS in this patient population by obtaining blood and urine Table 3 is a summary of the clinical toxicities seen during the samples at a fixed dose. A total of 14 patients were treated at this trial across all dose levels.

Table I Patient characteristics Table 3 Toxicity”

Characteristic Value Grade 21 Total no. of patients Toxicity I 2 3 4 Age (yr) Median 64 Nausea/vomiting 1 Range 34-75 Fatigue I I Sex (no. of males/no. of females) I 1/10 Myalgia I Race (no. of caucasians/no. of 19/2 Dizziness 2 African-Americans) Diarrhea 2 4 Performance status Stomatitis I 0 11 Proctitis 2 5 9 GI bleeding 16” 4’ 2 Irritability I Cancer diagnosis Liver enzymes I Sarcoma 6 Anemia 1 Colon S Hematuria 6 3 2 3 Ovary “ Toxicity was determined by National Cancer Institute Common Non-small cell lung 2 Toxicity Criteria, except in the case of GI bleeding.

Breast 1’ Hemoccult-positive stools.

Head and neck ( Gross bleeding, not requiring acute transfusion. Prostate Thyroid Pancreas Previous treatment bright red blood in her stools. This symptom resolved within a No prior treatment 4 week following cessation of the PPS. Patient 6. who had a gross One prior treatment 3 More than one prior treatment 14 total resection of a pelvic recurrence of a leiomyosarcorna just prior to starting PPS, was on treatment for 7 months without evidence of disease progression. During the last 2 months of treatment, she developed increasing frequency and sense of Table 2 Dose levels urgency of bowel movements and hemoccult-positive stools. Dose levels (mg/m2) No. of patients treated Proctitis was diagnosed endoscopically and pathologically, and

180 4 she elected to discontinue the study. Despite having negative 270 3 staging evaluation at the time of removal from the study, she 400 14 was found to have tumor recurrence of 8 X 6 X 4 cm 3 months later. Patient 7 (metastatic non-small cell lung) was treated for 4 months without evidence of progression but developed hemoccult-positive stools and proctitis during month 4 of ther- Dose Level 1 (180 mg/m2). Four patients were treated on apy and elected to discontinue the PPS.

dose level I . Patient 1 was replaced because of her inability to Dose Level 3 (400 mg/m2). A total of 14 patients were complete more than 1 week of treatment, secondary to bowel treated on dose level 3, 6 during the initial phase of the trial and obstruction that was caused by the rapid progression of disease. 8 additional patients as outlined above. As the time accrual to Patient 2 (metastatic leiomyosarcorna) was treated for 28 weeks this dose level began, no patients had yet been identified as but was taken off study when she was found to have bloody, having dose-limiting GI toxicity. Three of the six patients in the mucous-containing stools and evidence for progressive disease. initial phase developed hernoccult-positive stools within the first Patient 3 was treated for 60 weeks with stable disease through- month of treatment, with another developing hemoccult-positive out this time. This patient had metastatic hemangiosarcoma in stools in month 3. Two patients developed grade 1 hematuria as his liver and spleen that had been progressing steadily prior to well. No evidence of antitumor activity or prolonged stabiliza- starting PPS. During this period, PPS was held for 1-2 weeks on tion was seen in this cohort. The rapid onset of the lower GI three different occasions, secondary to either heme-positive symptoms of urgency and hemoccult positivity. without effec- stools or symptoms of proctitis. Endoscopic evaluations were tive treatment other than stopping the agent. was also then noted positive only for the distal mucosal ulcers seen in other patients to occur in patients treated at the lower dose levels after more treated with PPS, and recovery was complete (as documented by prolonged exposure to PPS. repeat endoscopy) in <2 weeks. Following removal from the Eight additional patients were then treated at 400 mg/rn2 to study for persistent proctitis and bright red bleeding per rectum, better characterize the lower GI toxicities and to obtain phar- despite attempts at dose reductions and topical rectal steroids, macokinetic data from serum and urine to determine the effects his disease progressed rapidly, with both the emergence of new of PPS on bFGF levels in patients with advanced cancer. As lesions and growth of existing lesions. Patient 4 (metastatic seen in the original six patients, symptoms of proctitis were prostate) was treated for 3 months and was discontinued because common early in the course of treatment, and no antiturnor of progressive disease. activity was seen. All but two of the patients discontinued Dose Level 2 (270 mg/rn2). Three patients were treated treatment between days 28 and 56 for progressive disease. Of on dose level 2. Patient S (leiomyosarcorna) progressed after 8 those two treated for a longer period, one remained on study for weeks on treatment and developed lower GI bleeding, with 3 months and then was found to have progressive disease, and

Table 4 GI toxicity by dose level 600 Time to onset of GI toxicity by dose level (months)

Toxicity Dose Level I Dose Level 2 Dose Level 3 -J (all grades) ( I 80 mg/m2) (270 mg/m2) (400 mg/m2) E 400 a) Diarrhea 4 1, 3, 4 1. 1, 1, 1. 1. 2 C ci GI bleeding 1, 7 2. 4. 8 1, 1, I, I , I. 3 C

Proctitis 2. 7 2, 4, 6 1, I . I .“ 3” 8 a Clinical diagnosis only. E U) 200 a. one was without evidence of disease at the start and discontinued treatment, secondary to gastritis and proctitis. 0 Proctitis As discussed above, proctitis was the DLT in 0 4 8 12 16 20 24 this trial of p.o. PPS. Patients at the higher dose levels developed Time (hi) symptoms more rapidly than the lower levels, but proctitis was Fig. I Data points. mean plasma concentrations of PPS in cancer present (evident either clinically or endoscopically) in patients patients following single and multiple doses of PPS (400 mg/rn2): bars, at all levels (Table 4). The patients were evaluated with colono- SE. #{149}, day 1: 0. day 15. scopic examination, with the finding that punctate ulcers were found virtually exclusively in the distal rectum (only patient 6 had ulcers 12 cm above the anal verge). Biopsies of the ulcers Table 5 Summary of pharmacokinetic parameters for PPS” showed only nonspecific inflammation. Repeat endoscopies I week after stopping PPS showed healing of the areas. In those Ratio Day 1 Day 15 (day 20:day I) patients rechallenged with PPS. symptoms of proctitis returned within I week. Attempts to alter the treatment schedule (drop- Cmax (ng/ml) 68.7 ± 6.8 462 ± 240.8 6.7 ± 3.3 AUCh (ng . lJm1) 987.5 ± 358.7 6660.0 ± 3898.7 7.7 ± 4.7 ping the evening dose, changing to bid. dosing. and lowering Tm..x (h) 8.6 ± 10.8 4.3 ± 2.4 2.9 ± 3.6 the dose) failed to relieve the symptoms. The use of enemas (‘ a = 4: all patients were treated at 400 mg/m2. containing corticosteroids was marginally helpful but did not I, AUC. area under the concentration curve. completely relieve the symptoms. PPS Pharmacokinetics. In four of the patients at the

400-mg/rn2 dose, serum samples were taken at 0. 0.5. 1 . 2, 4, 6, Table 6 Urine bFGF levels (in pg/ml) following PPS 8, and 24 h after the morning dose on days I and 15. Concen- (at 400 mg/m2) trations of PPS were determined using the methodology described above. The linear portion of the assay ranged from 4 to Patient no. Cancer diagnosis Day 1 Day 15 60 ngml. No cross-reactivity was found with other exogenous 17 Colon 2899 6160 glycosaminoglycans, but there was cross-reactivity to another 18 Breast 94 479 19 Colon 1427 760 form of commercially available PPS. 20 Colon 158 408 Mean plasma concentrations of PPS from the four patients in whom we have complete data sets after administration of 400 mg/m2 PPS on days 1 and 15 are shown in Fig. 1 . As noted, levels were much higher after repeated administration compared bFGF levels rose significantly from day 1 to day 15 in the other to the first dose. Pharmacokinetic parameters obtained after the six patients tested. day 1 and day 15 doses are listed in Table 5. Marked accurnu- Serum bFGF. Serum samples that were obtained for lation of PPS (or species recognized by the antisera), as meas- pharmacokinetic analysis at time points 0, 0.5, 1, 2, 4, 6, 8, and ured by Cmax and area under the concentration curve ratio for 24 h were also analyzed for bFGF levels. The results from the days 1-15. were nearly 7- and 8-fold higher after multiple four patients with complete data sets treated at dose level 3 are dosing. The elimination half-life of PPS could not be ascertained shown in Fig. 2. On both day 1 and day 15, patients took one because samples were collected for only 24 h postdose. which dose of PPS at time 0, without repeating the dose until the was inadequate, given the slow rate of terminal clearance of PPS. following day. However, patients had taken a dose of PPS 6 h Urine bFGF. Twenty-four-h urine samples were ob- prior to the day 15 dose. Although there is a suggestion that tamed from seven of the eight patients in the additional cohort bFGF fell precipitously (within 1 h) following a single dose of on days 1 and 15. Aliquots were obtained from these samples P.O. PPS (patients 17 and 20), the data, as a whole, fail to and analyzed for bFGF levels. Results from the four patients in demonstrate a consistent, interpretable pattern. Day 15 levels whom all samples were obtained sets are shown in Table 6. remained low, rising only when PPS was held for the remainder bFGF levels in both urine and serum were determined using a of the 24-h period. Patient 19 remains the exception, as little human FGF basic ELISA kit (R&D Systems). Interference effect was seen on bFGF concentrations in serum or urine in assays were run, which showed that PPS itself was not interfer- relation to PPS administration, despite the attainment of corn- ing with the analysis. With the exception of patient 19, urine parable serum levels of PPS.

Day #1 Day #15 16 16

14 14

12 12 Fig. 2 Serum bFGF levels. Se- 10 -. 10 rum samples were obtained at E . time points 0, 0.5, 1, 2, 4. 6, 8, 8 and 24 h. Data points, results U. U. from the four patients with com- 0 6 06 U. U. plete data sets treated at dose 4 4 level 3; bars, SD. #{149}, patient 17: EL patient 18; A. patient 19: V. 2 2 patient 20.

0 0

0 2 4 6 8 1012141618202224 0 2 4 6 8 1012141618202224

Time (hrs) ‘rime(hrs)

Relationship of Serum bFGF to PPS. When all seven absorbed PPS binding to mucosal bFGF, depleting the local patients were included in the analysis, there was no significant tissues of a potential protectant and/or repair stimulant. We have

correlation between bFGF and PPS (P = 0.85). However, this concluded that PPS is unacceptable for p.o. chronic administra- null result was heavily influenced by a single patient (Patient tion at doses of l80 mg/rn2 t.i.d. 21), who had two large jumps in both bFGF and PPS at 1 and We were not surprised at the lack of complete or partial 24 h. When the analysis was repeated with patient 21 omitted, responses in this trial, given the presumed antiangiogenic mech- there was a significant negative correlation between bFGF and anism of action of PPS. The only evidence for antitumor activity PPS (P = 0.02). Each unit increase in PPS level (in ng/rnl) was in this trial came from three patients with previously progressing associated with a 0.046 (95% confidence interval, 0.008-0.083) disease, who were maintained on PPS for at least 6 months. drop in bFGF level (in pg/rnl). Following discontinuation of PPS, tumor growth resumed within 1-3 months in these patients. All three patients suffered DISCUSSION from advanced sarcomas which, although relatively indolent, Here, we have determined that PPS can be administered eventually progressed and led to the death of one patient 3 safely to patients with advanced cancer without significant months after PPS was stopped. Sarcomas, as a class, are corn- systemic toxicity (apart from proctitis) at a dose of 400 mg/rn2 monly felt to be highly vascular tumors and logical choices in P.O. t.i.d. for a limited time period but that it is not tolerable for which to explore antiangiogenic agents. prolonged dosing. The DLT with chronic dosing was rectal A unifying explanation of PPS’s action in V1VO may be ulcerations, which reversed within approximately 1 week fol- suggested from the data obtained in this trial. There was an lowing cessation of the agent. No other significant toxicities inverse relationship between PPS and bFGF levels in the serum were seen. This finding is in contrast to preexisting data, show- and a direct relationship between PPS levels in the serum and ing p.o. PPS to be effective palliation for radiation-induced bFGF levels in the urine. Our hypothesis is that PPS binds bFGF proctitis, and may suggest that the ulcers are time and dose and is subsequently cleared by the kidney. The urine bFGF related because the MTD used in the proctitis trial was 300 mg levels show a consistent pattern of bFGF increase with contin- (not per m2) t.i.d. and was given for no longer than 3 months ued exposure to PPS (except in patient 19), and the serum (37). Here, the ulcers were seen in all dose levels, but they pharmacokinetic analysis shows an accumulation of PPS in the occurred more rapidly as the dose was increased, again suggest- body with chronic administration. This latter observation has ing a relationship to total dose administered. Rectal ulcers were been duplicated in normal volunteers (data not shown). A po- not noted in either the iv. or s.c. PPS trials. Pathological and tential fault of the protein kinase data is that we are unsure if we endoscopic evaluations failed to point toward a cause for this are measuring the active component of PPS because it is a large, finding. One possible mechanism for the generation of the ulcers polymorphic molecule. We failed to show any antiangiogenic may lie in the need for a high concentration of sulfur-containing activity from the serum of patients treated during this trial in our side chains in the mucous secreted by the rectum to serve as an bioassay, which suggests that the level of detection of the effective mucosal barrier. p.o. bioavailability data suggest that bioassay is too low, the levels of PPS were below the threshold most of the ingested PPS is not absorbed in the GI tract and is antiangiogenic activity, or what we are measuring in the serum therefore concentrated in the distal rectum prior to evacuation. by RIA is inactive. Unabsorbed PPS may serve as a sulfur sink, removing it from Continued preclinical work with PPS in the laboratory has the normal mucous and altering its ability to protect the under- shown that PPS is a more effective inhibitor of tumor growth in lying mucosa. Another possible mechanism could relate to un- rats when they receive it in a weekly i.p. schedule rather than

being exposed to it continuously (pattern seen at all dose levels); 16. Brem, S.. Gullino, P., and Medina, D. Angiogenesis: a marker for this is the strongest evidence to date showing the anticancer neoplastic transformation of mammary papillary hyperplasia. Science (Washington DC), 195: 880-882, 1977. activity of PPS. With continued exposure to PPS and subsequent depletion of bFGF, cells that produce bFGF may be up-regu- 17. Jensen. H. Angiogenesis induced by normal human breast tissue. In.- D. Rifkin and M. Klagsbrun (eds.). Angiogenesis: Mechanisms and lated to produce more bFGF in response. This is contrary to the Pathology. Current Communications in Molecular Biology, pp.155-157. commonly held belief that resistance is unlikely to be seen when Cold Spring Harbor. NY: Cold Spring Harbor Laboratory. 1987. an antiangiogenic agent is used. Support for this hypothesis 18. Weidner, N., Semple, J. P.. Welch, W. R., and Folkman, J. Tumor comes from our results, which show a rise in the urine bFGF angiogenesis and metastasis: correlation in invasive breast carcinoma. levels. (An alternate hypothesis is that the accumulation of PPS N. Engl. J. Med., 324: 1-8, 1991. in the body and tissues extracts more bFGF, which is subse- 19. Maeda, K., Chung, Y., Takatska, 0., Ogawa, Y.. Sawada, T., quently excreted.) The intermittent schedule of administration Yamashita. Y.. Onada, N., Kato, Y., Nitta, A., and Arimoto, Y. Tumor angiogenesis as a predictor of recurrence in gastric carcinoma. J. Clin. we described in rats may allow for the “washing out” of bFGF Oncol., 13: 477-481, 1995. from the serum and interstitium and prevent the up-regulation of 20. Maciag, T., Hoover, G., Stemerman, M., and Weinstein, R. Serial bFGF secretion. Future preclinical and clinical trials will at- propagation of human endothelial cells in vitro. J. Cell Biol., 91: tempt to answer these questions. 420-426, 1981. 21. Shing. Y.. Folkman, J., Haudenschild, C., Lund, D., Crum, R., and Klagsbrun, M. Angiogenesis is stimulated by a tumor-derived endothe- REFERENCES hal cell growth factor. J. Cell. Biochem., 29: 275-287, 1985. 1 . Folkman. J. What is the evidence that tumors are angiogenesis 22. Thomas, K., Rios-Candelore, M., Gimenez-Gallego. G., DiSalvo, dependent? J. Natl. Cancer Inst. (Bethesda), 82: 4-6. 1990. J., Bennett, C.. Rodkey, J., and Fitzpatrick, S. Pure brain-derived acidic 2. Algire, G.. Chalkley, H., Legallais, F., and Park, H. Vascular reac- fibroblast growth factor is a potent angiogenic vascular endothelial cell tions of normal and malignant tumors in vis’o. I. Vascular reactions of mitogen with sequence homology to interleukin 1 . Proc. NatI. Acad. Sci. mice to wounds and to normal and neoplastic transplants. J. NatI. Cancer USA, 82: 6409-6413, 1985. Inst. (Bethesda), 6: 73-85. 1945. 23. Esch, F., Baird, A., Ling. N.. Ueno. N., Hill, F., Denoroy. L., 3. Gimbrone, M., Leapman. S.. Cotran. R., and Folkman, J. Tumor Klepper. R.. Gospodarowicz, D., Bohlen, P., and Guillemin, R. Primary dormancy in rico by prevention of neovascularization. J. Exp. Med., structure of bovine pituitary basic fibroblast growth factor (FGF) and /36: 261-276, 1972. comparison with the amino-terminal sequence of bovine brain acidic 4. Knighton, D., Ausprunk, D., Tapper, D.. and Folkman, J. Avascular FGF. Proc. Nail. Acad. Sci. USA, 82: 6507-651 1, 1985. and vascular phases of tumour growth in the chick embryo. Br. J. 24. Lobb. R., Alderman, E., and Fett, J. Induction of angiogenesis by Cancer, 35: 347-356, 1977. bovine brain-derived class I heparmn-binding growth factor. Biochem- 5. Nicosia, R., Tchao, R., and Leighton, J. Angiogenesis-dependent istry, 24: 4969-4973, 1985. tumor spread in reinforced fibrin clot culture. Cancer Res., 43: 2159- 25. Davidson, J., Klagsbrun, M., Hill, K., Buckley, A., Sullivan, R., 2166, 1983. Brewer, P.. and Woodward, S. Accelerated wound repair, cell prolifer- 6. Folkman, J. Angiogenesis and its inhibitors. In: V. DeVita, S. Hell- ation, and collagen accumulation are produced by a cartilage-derived man, and S. Rosenberg (eds.). Important Advances in Oncology 1985, growth factor. J. Cell Biol., 100. 1219-1227. 1985. pp. 42-62. Philadelphia: J. B. Lippincott Co., 1985. 26. Terranova, V., DiFlorio, R., Lyall, R., Hic. S., Friesel, R., and 7. Liotta, L.. Kleinerman, J., and Saidel, G. The significance of hema- Maciag. T. Human endothelial cells are chemotactic to endothelial cell togenous tumor cell clumps in the metastatic process. Cancer Res.. 36: growth factor and heparin. J. Cell Biol., 101: 2330-2334, 1985. 889-894. 1976. 27. Sato. Y., and Rifkin, D. Autocrine activities of basic fibroblast 8. Liotta, L., Kleinerman, J., and Saidel, G. Quantitative relationships growth factor: regulation of endothelial cell movement, plasminogen of intravascular tumor cells, tumor vessels, and pulmonary metastases activator synthesis, and DNA synthesis. J. Cell Biol.. 107: 1199-1205, following tumor implantation. Cancer Res., 34: 997-1004, 1974. 1988. 9. Nagy, J., Brown, L., Senger, D.. Lanir, N., Van Dc Water, L., 28. Thompson. J., Haudenschild. C.. Anderson, K.. DiPietro. J.. Ander- Dvorak, A., and Dvorak, H. Pathogenesis of tumor stroma generation: a son, W. F., and Maciag, T. Heparmn-binding growth factor 1 induces the critical role for leaky blood vessels and fibrin deposition. Biochim. formation of organoid neovascular structures in vito. Proc. NatI. Acad. Biophys. Acta, 948: 305-326, 1988. Sci. USA, 86: 7928-7932, 1989. 10. Folkman, J., Watson, K., Ingber, D., and Hanahan, D. Induction of 29. Burgess. W., and Maciag, T.. The heparin-binding (fibroblast) angiogenesis during the transition from hyperplasia to neoplasia. Nature growth factor family of proteins. Annu. Rev. Biochem., 58: 575-606, (Lond.). 339: 58-61, 1989. 1989. I 1. Sillman, F.. Boyce. J.. and Fruchter. R. The significance of atypical 30. Baird, A., and Klagsbrun, M. The fibroblast growth factor family. vessels and neovascularization in cervical neoplasia. Am. J. Obstet. Gynecol., /39: 154-159, 1981. Cancer Cells, 3: 239-243, 1991. 12. Srivastava, A., Laidler, P., Hughes, L.. Woodcock, J.. and Shedden, 31. Folkman, J., and Klagsbrun. M. Angiogenic factors. Science E. Neovascularization in human cutaneous melanoma: a quantitative (Washington DC). 235: 442-447, 1987. morphological and Doppler ultrasound study. Eur. J. Cancer Clin. On- 32. Rifkin, D.. and Moscatelli, D. Recent developments in the cell col., 22: 1205-1209, 1986. biology of basic fibroblast growth factor. J. Cell Biol.. 109: 1-6. 1989. 13. Chodak, G., Haudenschild, C., Gittes, R., and Folkman, J. Angio- 33. Wellstein, A., Zugmaier, G., Califano, J., Broder, S., and Lippman, genic activity as a marker of neoplastic and of preneoplastic lesions of M. Xylanpolyhydrogensulfate inhibits fibroblast growth factor-depen- the human bladder. Ann. Surg., 192: 762-771, 1980. dent growth of human tumor cells. Proc. Am. Assoc. Cancer Res., 30: 14. Jensen, H., Chen, I., DeVault, M., and Lewis, A. Angiogenesis 583, 1989. induced by “normal” human breast tissue: a probable marker for pre- 34. Wellstein, A., Zugmaier. G., Califano, J., Kern, F., Paik, S., and cancer. Science (Washington DC). 218: 293-295, 1982. Lippman, M. Tumor growth dependent on Kaposi’s sarcoma-derived 15. Brem, S., Jensen, H., and Gullino, P. Angiogenesis as a marker for fibroblast growth factor inhibited by pentosan polysulfate. J. NatI. preneoplastic lesions of the human breast. Cancer (Phila.), 4/ 239-244, 1978. Cancer Inst. (Bethesda), 83: 716-720. 1991.

35. Wagenvoord, R., Hendrix, H., Soria, C., and Hemker, H. Localiza- 42. MacGregor, I., Dawes, J., Pepper, D., Prowse, C., and Stocks, J. tion of the inhibitory sites of pentosan polysulfate in blood coagulation. Metabolism of sodium pentosan polysulphate in man measured by a new Thromb. Haemostasis, 60: 220-225, 1988. competitive binding assay for sulphated polysaccharides: comparison 36. Tardy-Poncet, B., Tardy, B., Grelac, F., Reynaud, J., Mismetti, P., with effects upon anticoagulant activity, lipolysis, and platelet a-granule Bertrand, J. C., and Guyotat, D. Pentosan polysulfate-induced throm- proteins. Thromb. Haemostasis, 53: 41 1-414, 1985. bocytopenia and thrombosis. Am. J. Hematol., 45: 252-257, 1994. 43. Dawes, J., Prowse, C., and Pepper, D. The measurement of heparin 37. Grigsby, P. W., Pilepich, M. V., and Parsons, C. L. Preliminary results and other therapeutic sulphated polysaccharides in plasma, serum, and of a Phase I/LI study of sodium pentosan polysulfate in the treatment of urine. Thromb. Haemostasis, 54: 630-634, 1985. chronic radiation-induced proctitis. Am. J. Clin. Oncol., 13: 28-31, 1990. 44. Dawes, J., Prowse, C., and Pepper, D. Absorption of heparin, LMW 38. Mulholland, S., Hanno, P., Parsons, C., Sant, G., and Staskin, D. heparin. and SPS4 after subcutaneous injection, assessed by competitive Pentosan polysulfate sodium for therapy of interstitial cystitis: a double- binding assay. Thromb. Res., 44: 683-693, 1986. blind, placebo controlled clinical study. Urology, 35: 552-558, 1990. 45. Callahan, H. J., Shupp-Byrne, D., Pizzo, M., Parsons, C. L., and 39. Parsons, C. L., and Mulholland, S. G. Successful therapy of inter- Mulholland, S. G. The production of antibodies to pentosanpolysulfate stitial cystitis with pentosanpolysulfate. J. Urol., 138: 513-516, 1987. (ELMIRON, SP-54). J. Immunol. Methods, 136: 53-59, 1991. 40. Danielson, B., Fellstrom, B., Lindsjo, M., Ljunghall, S., and Wikstrom, 46. Liang, K-Y., and Zeger, S. L. Longitudinal data analysis using B. New drugs to prevent recurrence of renal stone disease. In: Proceedings of 11th International Congress of Nephrology, Tokyo: 1990. generalized linear models. Biometrika, 73: 13-22, 1986. 41. Swain, S. M., Parker, B., Wellstein, A., Lippman, M. E., Steakley, C., and DeLap, R. Phase I trial of pentosan polysulphate. Invest. New Drugs, 13: 55-62, 1995.

Downloaded from clincancerres.aacrjournals.org on September 24, 2021. © 1997 American Association for Cancer Research. Phase I trial of orally administered pentosan polysulfate in patients with advanced cancer.

J L Marshall, A Wellstein, J Rae, et al.

Clin Cancer Res 1997;3:2347-2354.

Updated version Access the most recent version of this article at: http://clincancerres.aacrjournals.org/content/3/12/2347

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/3/12/2347. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.