Phase I Study of the Novel Cyclic AMP (cAMP) Analogue 8-Chloro-cAMP in Patients with Cancer: Toxicity, Hormonal, and Immunological Effects

One of the major signal transduction pathways in mammalian cells is the cAMP²-dependent pathway (1, 2). cAMP acts by binding to the regulatory subunits of PKAs. There are two main classes of PKA, PKAI and PKAII, which share identical catalytic but different regulatory subunits (3). These regulatory subunits are termed RI and RII. PKAI and/or its RI regulatory subunit is overexpressed in cancer cell lines and tumor specimens and up-regulated during growth factor-mediated cell transformation (4, 5). In addition, RI antisense oligonucleotides cause growth arrest in a variety of human cancer cell lines (6, 7) and in vivo (8). Hence, decreased RI:RII ratios and/or up-regulation of PKAII correlate with growth inhibition and cellular differentiation (9).

8-Cl-cAMP is a cAMP analogue that preferentially binds to RII. It induces rapid RII up-regulation and eventual RI down-regulation (10, 11). It has potent inhibitory effects on a wide variety of human cancer cell lines, with an IC₅₀ ranging from 0.1 to 20 μm(12). The IC₅₀ falls with length of drug exposure (13). 8-Cl-cAMP can suppress c-myc and c-ras proto-oncogenes in vitro (14, 15) and in vivo (10) and induce differentiation in transformed human cancer cell lines and maturation of leukemic blasts obtained from patients with leukemia (reverse transformation; Refs. 9 and 16). 8-Cl-cAMP does not significantly inhibit the growth of NIH 3T3 cells, rat kidney fibroblasts, mammary epithelial cells, or peripheral blood lymphocytes (17), nor does it inhibit the growth of parental cells whose progeny have been transformed (11). Thus, it inhibits growth selectively in transformed rather than nontransformed cells. These properties make it an attractive candidate for cancer therapy because they suggest that it should not cause the toxicity of conventional cytotoxic agents but should inhibit tumor growth. In addition, 8-Cl-cAMP can down-regulate mdr-1 expression (13) and synergize with conventional cytotoxic agents (18, 19).

Dogs receiving 8-Cl-cAMP (0.34–6.67 mg/kg/h) by continuous infusion developed gastrointestinal and renal toxicity, even at the lowest dose administered (20, 21). These toxicities required the treatment to be halted by day 25 at the lowest dose to prevent fatalities. Hematological toxicity was not observed. In view of the possibility that 8-Cl-cAMP could act specifically on cancer cells without affecting normal cells, we initiated a Phase I study of 21-day continuous infusion 8-Cl-cAMP in patients with cancer, aimed at assessing the drug’s toxic, pharmacokinetic, and pharmacodynamic properties.

PKAs are involved in hormone-dependent signal transduction (22, 23). 8-Cl-cAMP may modulate these signals. To assess pharmacodynamic end points, we determined the effect of 8-Cl-cAMP on circulating basal levels of representative pituitary, thyroid, adrenal, and gonadal hormones. In addition, as a dynamic test for effects on post-hormone receptor events, we measured the effect of 8-Cl-cAMP on the cortisol, androstenedione, and 17-hydoxyprogesterone response to adrenocorticotropic hormone.

PKAs are also involved in early events in lymphocyte activation, including IL-6 and TNF-α gene transcription (24, 25, 26). Any changes associated with 8-Cl-cAMP treatment were, therefore, determined for circulating IL-6 and TNF-α levels and for markers of peripheral blood lymphocyte activation, as a further assessment of any pharmacodynamic effects.

Patients had a histological diagnosis of cancer and an Eastern Cooperative Oncology Group performance status of 0–2, and they either had failed conventional therapy or had tumors for which there was no standard treatment (Table 1). A complete history, full examination, chest X-ray, and 12-lead electrocardiogram were obtained prior to entry into the study. Patients were required to have a white cell count of >3.0 × 10⁹/liter, platelet count of >100 × 10⁹/liter, normal renal and hepatic function, and a life expectancy of at least 3 months and to give written informed consent. Patients had not received radiotherapy or chemotherapy in the 4 weeks (6 weeks for nitrosoureas or mitomycin C) before the start of the study. The study was approved by the Central Oxford Research Ethics Committee and conducted according to the Declaration of Helsinki.

8-Cl-cAMP was supplied as the sodium salt by Dr. K. Miki (Fundamental Research Laboratories, Tonen Corporation, Kanagawa, Japan), with the exception of a single patient treated at the highest dose level, for whom it was provided by ICN Pharmaceuticals Inc. (Belgrade, Yugoslavia). The drug was dissolved in 5% dextrose and infused via a Hickman line, using a Walkmed 350 ambulatory infusion pump (Medfusion Inc., Duluth, GA). Each patient remained an in-patient for the first 24 h of the first cycle of treatment, after which treatment was given on an outpatient basis.

The drug was initially given by continuous infusion for 21 days, followed by 1 week of rest. This 4-week period was considered one treatment cycle (Table 2). Dose level 1 was 0.005 mg/kg/h, a dose ∼50 times lower than that which caused toxicity in preclinical studies. Escalation to 0.045 mg/kg/h was achieved in three increments (Table 2). At least two patients received the drug at each dose level. During the study, it became evident that a continuous infusion for 3 weeks per month at the highest dose level was not possible because of hypercalcemia. The schedule was, therefore, changed to 5 days per week for 3 weeks, followed by 1 week of rest (Table 2). Subsequently, it was further modified to 3 days per week for 3 weeks, followed by 1 week of rest.

The MTD was defined using National Cancer Institute Common Toxicity Criteria and was defined as: at least three of six patients with grade 3 hematological toxicity or two of six patients with grade 4 hematological toxicity or two of six patients with grade 3 nonhematological toxicity.

Patients were reviewed by a physician weekly, and new signs, symptoms, and performance status (Eastern Cooperative Oncology Group) were documented. At each visit, full blood counts with a differential white cell count, serum biochemistry, and urinalysis were performed. Additional tests were performed as appropriate. Patients continued therapy if there was no evidence of disease progression after one cycle of treatment and if there was acceptable or no toxicity. Although this was a Phase I study, patients were monitored for evidence of disease response.

Evaluable and measurable disease sites were assessed before entering the study by physical examination, plain radiography, and computerized tomography, where appropriate, and repeated every two cycles. Physical examination was repeated weekly and imaging investigations for the purposes of tumor measurement were repeated after two cycles of treatment or at the time of suspected disease progression.

Standard WHO criteria for objective response assessment were used. Partial response was defined as a ≥50% reduction in the sum of the products of the largest perpendicular diameters of all measurable disease sites. Progressive disease was indicated by a ≥25% increase in the size of at least one measurable lesion or the appearance of a new lesion. Stable disease was defined as an increase in disease measurements of <25% or a decrease of <50%.

In three patients receiving 0.045 mg/kg/h, venous blood was obtained 0, 5, 10, 15, 30, 45, 60, 90, 120, 150, 180, 210, and 300 min after discontinuation of the first infusion. It was collected into EDTA tubes, spun at 2000 rpm, and aliquoted, and the supernatant was stored at −70°C. Analysis was by reversed-phase high-performance liquid chromatography, as described previously (27).

CD4, CD8, CD20 (B cells), CD23 (activated B cells), CD45RA (“naive” T cells), CD45R0 (“memory” T cells), and natural killer cells were measured by dual-color flow cytometry, by previously described methods (28), in six patients receiving 0.045 mg/kg/h (5-day infusion) on days 0, 5, and 20 of the first treatment cycle. In eight patients receiving the same dose and schedule, plasma IL-6 and TNF-α levels were measured at the same time points, by RIA and by immunoradiometric assay, respectively, as described previously (29).

The following serum hormone levels were assayed, by previously described methods (30), during the first treatment cycle: cortisol, androstenedione, estradiol, testosterone, 17-hydoxyprogesterone, LH, follicle-stimulating hormone, prolactin, sex hormone-binding globulin, and thyroid-stimulating hormone. These samples were obtained immediately before starting the infusion and on the last day of treatment in the first treatment cycle (day 20). In addition, the responses to 250 μg of s.c. tetracosactrin (synacthen) given 1 h before commencing treatment (day 0) and 1 h before drug withdrawal (day 20) at the end of the first cycle were measured for cortisol, androstenedione, and 17-hydoxyprogesterone. The blood samples were obtained immediately before and 0.5 h after synacthen administration.

Spot urine samples were collected for urinary NAGs at the start and end of drug administration. The samples were taken at the same time of day, and the first sample of the day was not used. Twenty ml of urine were spun at 2000 rpm, the supernatant was removed and placed in a fresh tube, frozen, and stored at −70°C for later analysis. NAGs were measured spectrophotometrically (31) using a commercial reagent kit (PPR Diagnostics, London, United Kingdom) and expressed as the amount of urinary NAG (in units) per mmol of urinary creatinine.

Student’s paired t tests were used for analysis of all data, and Ps < 0.05 were considered significant. The drug terminal elimination trend line was determined by the least squares method.

Thirty-two patients were treated. Their characteristics are shown in Table 1. Twenty-nine patients completed at least one cycle, and 22 completed at least two. There were four patients who received more than two cycles: two patients received 3 cycles before withdrawal for progressive disease; one patient received 6 cycles; and one patient received 10 cycles. Of the 29 patients who completed at least one cycle, 27 stopped treatment because of progressive disease, and 2 stopped because of toxicity. Three patients stopped treatment before completing one cycle. In one patient, this was because of toxicity; in one, it was because of progressive disease; and in one, it was because of both these events. Seven dose levels were used (Table 2).

The principal toxicity in this study was hypercalcemia (Table 2), which was the DLT. The hypercalcemia worsened with dose increments and was related to duration of infusion and dose. No hypercalcemia was observed at the first two dose levels. At the third dose level (0.025 mg/kg/h), of the three patients treated, one developed grade I toxicity, and another developed grade II toxicity. At the next dose level (0.045 mg/kg/h), the first two patients treated, who both received the drug continuously for 21 days, developed symptomatic hypercalcemia (grades I and III, respectively).

The length of treatment was therefore reduced from continuous infusion to 5 days per week for the first 3 weeks of a 4-week treatment cycle, and the dose was escalated to 0.08 mg/kg/h. Because one patient then developed grade IV hypercalcemia, the dose was reduced to 0.045 mg/kg/h. Nine patients were treated at this dose level and symptomatic hypercalcemia was again the DLT (Table 2). The drug administration time was, therefore, reduced to 3 days per week for the first 3 weeks of each 4-week treatment cycle, which enabled escalation to the top dose level of 0.15 mg/kg/h. At this level, the DLT was again symptomatic hypercalcemia (Fig. 2). Two patients developed grade III toxicity, and one patient developed grade II hypercalcemia.

Three of the four patients at the top dose level of 0.15 mg/kg/h also developed reversible hepatic toxicity, as indicated by two patients with grade II increases in bilirubin and one patient with grade III bilirubin increases (Table 2), one of whom also developed a grade I increase in aspartate transaminase.

Reversible and significant rises in creatinine were observed, generally in conjunction with hypercalcemia. The increase in creatinine only exceeded grade I toxicity in two patients, who both received 0.08 mg/kg/h, and returned to baseline values within a maximum of 7 days in all patients. Urinary NAG levels were measured in seven patients receiving 0.045 mg/kg/h (5-day infusion). Reversible increases in NAG levels were observed during the infusion, which reverted to baseline over 10 days (Fig. 1).

There was no significant hematological toxicity. The symptomatic toxicities were nausea, vomiting, and fatigue (Table 3), and these were generally noted at times of hypercalcemia. Increasing the daily fluid intake to 3 liters appeared to reduce the increase in creatinine and the symptomatic toxicities.

Three patients receiving 0.045 mg/kg/h over 5 days who developed hypercalcemia were treated with prophylactic i.v. disodium pamidronate at 60 mg per week. This did not significantly reduce the treatment-associated increase in calcium. Likewise, prophylactic use of oral dexamethasone (4 mg twice a day) in three patients receiving 0.045 mg/kg/h over 5 days failed to significantly reduce the drug-induced increase in calcium. Three patients receiving 0.15 mg/kg/h over 3 days were treated with salcatonin (salmon calcitonin) at 2–5 units/kg/day during the period of infusion, which also had no significant effect on the drug-induced calcium increase.

One mechanism of hypercalcemia may be through increased gastrointestinal calcium absorption. Hence, at the top dose levels of 0.11 and 0.15 mg/kg/h, patients who developed hypercalcemia during the first cycle of treatment received oral sodium cellulose phosphate (calcisorb), 500 mg three times daily, in the second cycle. The calcisorb was given during the time of 8-Cl-cAMP administration, to reduce to reduce gastrointestinal calcium absorption. The mean peak calcium levels in the second cycle of treatment were lower than in the first (Fig. 2), but not significantly so. Because of the small sample size, a significant modifying effect of calcisorb on calcium levels may have been missed.

Hormone levels were assayed in 16 patients receiving 0.045 mg/kg/h (5-day infusion), 0.08 mg/kg/h (5-day infusion), and 0.11 mg/kg/h (3-day infusion) during the first treatment cycle, immediately before starting treatment and on the last day of the treatment cycle. The mean androstenedione, estradiol, testosterone, thyroid-stimulating hormone, and basal cortisol levels at the start and end of treatment were not significantly different from each other. The mean end treatment LH was significantly higher than the mean start level (15.8 ± 4.2 versus 10.8 ± 3.2 IU/liter; P < 0.02). Likewise, the mean end basal 17-hydoxyprogesterone level was significantly higher than the mean start level (5.6 ± 0.74 versus 3.8 ± 0.7 nmol; P < 0.0006). The response to synacthen was measured for androstenedione, cortisol, and 17-hydoxyprogesterone, but there were no significant differences between either the peak or the incremental rise in any of these three hormones between the start and end of 8-Cl-cAMP treatment.

Because of the influence of sex on sex hormones, the data obtained for females and males were also separately analyzed. Mean 17-hydoxyprogesterone levels at the start and end of treatment were 1.8 ± 0.5 versus 4.0 ± 0.7 nmol (P < 0.003), respectively, for females and 5.1 ± 0.9 versus 6.7 ± 1.0 nmol (P < 0.04) for males. Mean LH levels at the start and end of treatment for males were 6.4 ± 0.9 versus 9.4 ± 1.8 IU/liter (P < 0.03), respectively. There was not a significant difference between mean LH levels at the start and end of treatment for females (16.5 ± 6.9 versus 24.1 ± 8.6 IU/liter), respectively. Separate analysis for premenopausal and postmenopausal females was not performed, because there were only two premenopausal females in this group.

Plasma IL-6 levels were measured in eight patients receiving 0.045 mg/kg/h (5-day infusion), and there were no significant changes on either day 5 or 20 of treatment compared with the baseline pretreatment levels. Likewise, analysis of peripheral blood lymphocyte subsets at the same time points showed no significant change with treatment. Plasma TNF levels were measured at the same time points but were undetected throughout the study.

Although this was a Phase I study, response data were obtained for the 29 patients who completed at least one cycle of treatment. There were no complete or partial responses. Twenty-seven patients had progressive disease. In 23 of these, this was apparent by the end of the first (7 patients) or second treatment cycle (16 patients). The other two patients (0.01 mg/kg/h over 21 days and 0.08 mg/kg/h over 5 days) had developed progressive disease by the end of the third treatment cycle. Two patients developed stable disease. One of these, with malignant thymoma, had progressive disease before treatment (0.045 mg/kg/h over 5 days), which remained stable for 40 months. The other patient (0.15 mg/kg/h over 3 days) with progressive colonic cancer attained stable disease for 4 months.

8-Cl-cAMP levels were measured to determine the terminal elimination pattern for 3 patients receiving 0.045 mg/kg/h (Fig. 3). Data concerning the time to steady state has already been determined for patients receiving 0.045 mg/kg/h (32), as have steady-state levels up to a dose of 0.25 mg/kg/h (33). The mean plasma 8-Cl-cAMP level immediately before the infusion was discontinued was 0.367 ± 0.08 μmol.

The principal aim of this study was to determine the MTD and the DLT for continuous infusion 8-Cl-cAMP. The MTD was 0.045 mg/kg/h, and the DLT was symptomatic hypercalcemia. In another study with breast cancer patients receiving continuous infusion 8-Cl-cAMP at either 0.0225 or 0.045 mg/kg/h for 28 days, steady-state drug levels of 0.15–0.72 μm were attained (32). The IC₅₀ for 8-Cl-cAMP is variously reported at 0.1–20 μm and is dependent on culture system and length of exposure to the drug (12, 13). Hence, it was important to determine whether higher doses of the drug could be used, and so the administration was modified and the drug given for 5 days every 3 weeks with a 1-week gap between each treatment cycle. Using this regimen, we found that the MTD was also 0.045 mg/kg/h and that the DLT was symptomatic hypercalcemia. Pharmacokinetic data show that the mean steady-state plasma drug level was likely to be 0.36 μmol for patients receiving this dose (Fig. 3). The infusion schedule was, therefore, shortened to 3 days every week for 3 of 4 weeks, and the MTD was 0.15 mg/kg/h. The DLT was again hypercalcemia and, also, reversible hepatotoxicity. Hence, this Phase I study has provided information on the MTD for continuous, 5-day, and 3-day intermittent 8-Cl-cAMP regimens. With the 3-day intermittent regimen, the total amount of drug administered per cycle at the MTD of 0.15 mg/kg/h was 32.4 mg/kg. This is ∼40% more than the total dose per cycle (22.7 mg/kg) at the MTD for the continuous regimen of 0.045 mg/kg/h.

In a more detailed study of 16 patients in this trial, treated at dose levels up to 0.045 mg/kg/h, we have shown that the mechanism of 8-Cl-cAMP-induced hypercalcemia was a parathyroid hormone-like effect, causing increases in 1,25-dihydroxyvitamin D₃ levels by up to 14 times the baseline values (30). This hypercalcemia was not reduced by coadministration of the drug with either disodium pamidronate or dexamethasone. Patients were, therefore, treated with calcisorb during 8-Cl-cAMP administration because it binds and reduces intestinal calcium absorption, thereby antagonizing intestinal parathyroid hormone-like effects. Comparison of peak calcium levels before and during calcisorb treatment failed to show a significant reduction with treatment. Nonetheless, this was a small cohort of patients; a significant effect would have been hard to detect, and mean calcium levels during calcisorb treatment were lower than those without it. In Phase II studies, calcisorb might prove effective in modulating cAMP-induced hypercalcemia. We also assessed the effect of calcitonin in three patients but failed to find a significant effect on calcium levels. Langdon et al. (34) have provided preliminary evidence that calcitonin could reduce weight loss and hypercalcemia in 8-Cl-cAMP-treated mice, although their observations did not reach significance, possibly because of the small sample size.

Tortora et al. (33) have also performed a Phase I study of 8-Cl-cAMP, in which the drug was given for 5 days per week for 2 weeks, followed by a 1-week rest period. Although hypercalcemia occurred, it was not as marked as in our study, never exceeding grade II toxicity. At the top dose administered, 0.25 mg/kg/h, the two patients treated had to be withdrawn after the first week of treatment because of grade IV renal impairment. At the lower dose of 0.125 mg/kg/h, they treated 15 patients, only one of whom developed hypercalcemia (grade I). The total dose per 3-week cycle for this level was 30 mg/kg. This contrasts with the patients in our study treated at the top dose of 0.15 mg/kg/h (total dose per 4 week cycle 32.4 mg/kg), who all developed hypercalcemia. It is not clear why the incidence of hypercalcemia was lower in the study by Tortora et al. study (33) because the drug supply was the same. Differences in scheduling probably do not account for it because, although, in the regimen of Tortora et al., the drug was given for 2 of 3 weeks, rather than 3 of 4 weeks as in this study, we observed hypercalcemia by the second treatment week, and it did not significantly worsen in the third week (Fig. 2). Cummings et al. (32) performed a pharmacokinetic study of six patients with breast cancer who received 8-Cl-cAMP at 0.045 mg/kg/h continuously for 28 days. This was not a Phase I dose escalation study, but some of these patients developed hypercalcemia; however, the incidence and severity of hypercalcemia were not reported (32). The difference between the two studies that have noted hypercalcemia and the study of Tortora et al. (33) are not readily reconcilable, although differing populations and diet could be contributing factors.

Reversible abnormalities in renal and hepatic function were observed in our study. In mice, it has been shown that 8-Cl-cAMP preferentially accumulates in the kidney and liver (35), which could account for this toxicity pattern. The hepatic abnormalities were indicative of hepatocellular damage and were rapidly reversible. Hepatic dysfunction has not been reported in animal or clinical studies of the drug (21, 33). Mild reversible impairment of renal function also occurred. Although hypercalcemia may have contributed, the observed reversible elevations in urinary NAGs indicate a degree of direct tubular toxicity. Corresponding toxicity was observed in dogs (21). The treatment was otherwise well tolerated, with few patients stopping because of toxicity.

Although this was a Phase I study, we also assessed tumor responses. No responses were observed. However, with this type of drug, disease stabilization may be a more realistic end point than disease responses. Two patients with progressive disease before treatment developed stable disease, perhaps suggesting evidence of cytostatic activity. This is similar to the experience of Tortora et al. (33), who observed clinical evidence of cytostatic properties.

PKA is critical in the transduction of hormone-dependent signals (23). Hence, we looked for changes in a range of hormone levels with treatment. Only LH and 17-hydoxyprogesterone levels were altered during drug treatment. Why these hormone levels were altered and others were not is not clear and may be related to tissue distribution because the RII regulatory subunit is variably expressed and differs in its subcellular localization (23). Variables such as tissue distribution, activities of phosphoprotein phosphatases, and non-cAMP pathways influence responses of a particular cell (22, 36). The elevations in LH and 17-hydoxyprogesterone levels indicate that these pathways may be predominantly regulated by the RII subunit, as is the PTH-like response (30).

cAMP is involved in cytokine dependent signaling, including IL-6 up-regulation and B-cell activation (24, 25, 37). There was no evidence that the drug modulated either plasma IL-6 or TNF levels or increased the numbers of activated B cells. Perhaps the absence of change in any lymphocyte subsets is not surprising, given the generally low level of detection of activated lymphocytes, even in individuals facing immune challenge. In fact, 8-Cl-cAMP could be predicted to down-regulate some aspects of immune activation because the RI isoenzyme is involved in lymphocyte activation (38) and some RII agonists down-regulate TNF-α and IL-6 cytokine production (26). Nonetheless, Tortora et al. (33) were able to demonstrate expansion in the population of IL-2 receptor-positive lymphocytes and in a subpopulation of NK cells in 8-Cl-cAMP-treated patients receiving similar doses to this study.

8-Cl-cAMP is a cytostatic drug in vitro (13). This study shows that it is safe, and three MTDs were achieved without significant hematological toxicity. In vitro 8-Cl-cAMP synergizes with conventional cytotoxic agents and reverses multidrug resistance (13, 18, 19). This study indicates safe doses of the drug that could be used in Phase II studies in combination with conventional cytotoxic agents that may not exacerbate their myelotoxicity. Whether 8-Cl-cAMP should be administered as a low-dose prolonged infusion or a higher intermittent dose is not certain. In vitro evidence suggests that the drug has prolonged inhibitory effects that gradually decline over 2–3 days after drug withdrawal (14), so an intermittent regimen would be feasible. The inhibitory effects take 3–4 days to build up (12), and the IC₅₀ falls linearly with increasing time of exposure to 8-Cl-cAMP (13). The synergy of 8-Cl-cAMP with conventional cytotoxic drugs appears to be abrogated if it is administered before cytotoxic drugs but preserved if administered after (19), suggesting that the drug should be given as an intermittent regimen, commencing immediately after cytotoxic administration. It was possible to administer higher total tolerable drug doses by giving the drug over 3 days rather than 5. Hence, we suggest that, in Phase II trials, 8-Cl-cAMP should be given as a 3-day intermittent weekly regimen at 0.11 mg/kg/h so that it could be given with concomitant conventional chemotherapy.