Effects of the Src kinase inhibitor saracatinib (AZD0530) on bone turnover in healthy men: A randomized, double-blind, placebo-controlled, multiple-ascending-dose phase I trial

Abstract

Src is a nonreceptor tyrosine kinase thought to be essential for osteoclast function and bone resorption. We investigated the effect of the orally available Src inhibitor saracatinib (AZD0530) on bone turnover in healthy men. The study was part of a randomized, double-blind, placebo-controlled multiple-ascending-dose phase I trial of saracatinib. Fifty-nine healthy men (mean age 34.6 years) were divided into five cohorts; four with 12 subjects and one with 11 subjects, and randomized within each cohort in the ratio 3:1 to receive a single dose of saracatinib or placebo, respectively, followed 7 to 10 days later with daily doses for a further 10 to 14 days. Dosing levels of saracatinib ascended by cohort (60 to 250 mg). Markers of bone turnover were measured predose and 24 and 48 hours after the initial single dose and immediately before and 24 and 48 hours and 10 to 14 days after the final dose. Data from 44 subjects were included in the analysis. There was a dose-dependent decrease in bone resorption markers [serum cross-linked C-telopeptide of type I collagen (sCTX) and urinary cross-linked N-telopeptide of type I collagen normalized to creatinine (uNTX/Cr)]. At a dose of 250 mg (maximum tolerated dose), sCTX decreased by 88% [95% confidence interval (CI) 84–91%] and uNTX/Cr decreased by 67% (95% CI 53–77%) from baseline 24 hours after the final dose. There was no significant effect on bone formation markers. There were no significant adverse events. We conclude that inhibition of Src reduces osteoclastic bone resorption in humans. Saracatinib is a potentially useful treatment for diseases characterized by increased bone resorption, such as metastatic bone disease and osteoporosis. © 2010 American Society for Bone and Mineral Research

Introduction

Src kinase is a nonreceptor tyrosine kinase and a member of the Src family of protein kinases. Src has multiple functions in the regulation of both normal and oncogenic cell processes, including proliferation, differentiation, survival, and angiogenesis. It is ubiquitously expressed at low levels in the majority of cell types and, in the absence of the appropriate extracellular stimuli, is maintained in an inactive conformation through the phosphorylation of Tyr 530 in the regulatory domain. Platelets, neural cells, and osteoclasts are the only normal cells known to express high levels of Src. Src expression and/or activity is elevated in a variety of solid tumors, where it appears to be associated with the processes of angiogenesis, tumor migration, and invasion.1

In the osteoclast, Src kinase plays a part in many of the signaling pathways responsible for osteoclast survival, motility, and activation, including RANKL- and MCSF-mediated activation, RANKL-mediated survival, and integrin-mediated motility and polarization of cells.2–4 The unexpected finding that the only phenotype in the Src^–/– mouse is osteopetrosis indicated that Src is an essential requirement for osteoclastic bone resorption.5 It appears that this is due to a lack of activation and failure to form a ruffled border rather than to effects on osteoclast formation or survival because the number of osteoclasts has been shown to be either unaltered or indeed increased in in vitro and in vivo studies. For example, the number of osteoclasts in bone from Src^–/– mice was reported to be twice that in wild-type mice.6 However, the osteopetrotic phenotype of these mice is probably not entirely due to failure of osteoclastic bone resorption. Marzia and colleagues 7 have shown that decreased Src expression may enhance osteoblast differentiation and bone formation. Src is therefore a potential target for the treatment of diseases characterized by elevated bone resorption such as metastatic bone disease and osteoporosis.

Saracatinib is a novel orally available ATP-competitive inhibitor of Src kinase. The IC₅₀ for Src is 2.7 and 4 nM, <4 nM, 5 nM, and 10 nM for cYes, Lck, Lyn, and cFyn, respectively.8 It also inhibits Abl kinase (IC₅₀ 30 nM) but has very little activity against other tyrosine and serine-threonine kinases such as VEGF-2, PDGF, FGFR, c-Kit, and Aur-3.8 Abl kinase, another ubiquitously expressed nonreceptor tyrosine kinase, has been implicated in cell proliferation, differentiation, and oxidative stress response. A partial penetrance (50%) of a severe osteoporotic phenotype is seen in the Abl^–/– mouse.9 Saracatinib is currently in clinical development as an antiprogressive and anti-invasive drug in the treatment of a number of solid tumors and is also being evaluated as an antiresorptive in the treatment of lytic bone metastases.

In vitro studies have shown that saracatinib has significant antiresorptive effects in bone. Mullender and colleagues 10 have shown that saracatinib inhibits bone resorption in vitro using rabbit osteoclasts cultured on bovine bone slices. The total resorption area was decreased by 63% and almost 95% at 0.1 and 0.5 µM, respectively, compared with controls owing to a reduction in both the number of resorption pits and the average size of the individual pits. Additionally, saracatinib (0.1 to 10.0 µM) reduces bone resorption in organ cultures of neonatal mouse calvariae prelabeled with ⁴⁵Ca in a dose-dependent manner.10 In a preliminary single-ascending-dose safety and tolerability study, we established that a single 1000 mg dose of saracatinib suppresses bone resorption, as assessed by serum cross-linked βC-terminal telopeptide of type I collagen (sCTX) and urinary cross-linked N-terminal telopeptide of type I collagen normalized to creatinine (uNTX/Cr) in healthy men.11

The aim of this study was to assess the effect of the Src inhibitor saracatinib on bone turnover as part of an ascending-multiple-dose, randomized, double-blind, placebo-controlled phase I trial in healthy men designed to assess the safety, tolerability, pharmacokinetics, and pharmacodynamics of saracatinib.

Materials and Methods

The trial was performed at the AstraZeneca Clinical Pharmacology Unit (CPU), Alderley Park, Macclesfield, UK. Healthy adult men aged 18 to 55 years were recruited from volunteers drawn from the AstraZeneca Healthy Volunteer Subject Panel. Inclusion criteria included normal medical examination, body mass index (BMI) between 18 and 30 kg/m², and never having smoked or, in those who had smoked previously, not having smoked (or used any other nicotine products) in the 6 months prior to enrollment. Recent users of nicotine products were not included to eliminate any potential confounding effects on phase I assessments. Exclusion criteria included a significant history of drugs of abuse or a positive test for drugs of abuse; concomitant use or use within the 3 months prior to the trial of drugs that are clinically significant substrates, inducers, or inhibitors of CYP3A4, which is required for the metabolism of saracatinib; and history or presence of gastrointestinal, hepatic, or renal disease or other conditions known to interfere with absorption, distribution, metabolism, or excretion of drugs.

Study design

The objective of this study was to investigate the effect on bone turnover of multiple daily oral dosing with saracatinib in healthy male volunteers. The study was part of an ascending-multiple-dose, randomized, double-blind, placebo-controlled phase I trial designed primarily to assess the safety, tolerability, pharmacokinetics, and pharmacodynamics of saracatinib, results of which will be published separately. Subjects were recruited sequentially by cohort. Each cohort was to comprise 12 subjects, and within each cohort, subjects were randomly assigned to treatment with saracatinib or placebo in a 3:1 ratio, respectively. The randomization schedule was generated by the Biostatistics Group at AstraZeneca, a separate schedule being generated for each cohort. Randomization numbers were allocated sequentially as subjects within a cohort became eligible for randomization. Subjects, investigators, nursing staff involved in study days, clinical research associates, and laboratory staff were unaware of the randomized treatment until all decisions on the evaluability of the data from all subjects had been made and documented.

The study design, indicating only the dosing regimen and time points of sample collection for bone turnover marker measurements, is shown in Figure 1. The doses of saracatinib administered were 60 mg (cohort 1), 125 mg (cohort 2), 250 mg (cohorts 3 and 4), and 185 mg (cohort 5). The decision to start dosing at 60 mg was based on the findings of a single-ascending-dose pilot study.11 In light of emerging data indicating that the half-life of the drug was longer than had been predicted originally, the dosing regimen was amended after cohort 1 to increase the duration of the washout from 7 to 10 days and the multiple-dosing phase from 10 to 14 days. The decisions to proceed to the next cohort and the size of the dose to be given were taken by the Safety Monitoring Committee after reviewing safety, tolerability, and pharmacokinetic data from previous doses within the trial and data from the previous single-ascending-dose study.11, 12

Figure 1. Study design. Solid arrows represent saracatinib doses; open arrows represent sampling times for bone turnover markers (fasting blood and second-morning-void urine). ^aCohort 4 and 5. ^bCohort 5.

Subjects were enrolled into the study up to 28 days before the first dose. They attended the CPU on the morning of day –1 for safety blood samples to be taken and returned the same evening. The initial single dose was given the following morning (day 1), and subjects remained resident in the CPU until 48 hours after this dose (morning of day 3). Throughout the remainder of the washout period (days 4 to 7 or day 10 for cohort 1 or all other cohorts, respectively), they visited the CPU daily for safety monitoring. They returned to the CPU the evening before the start of the multiple once-daily dosing phase (day 7 or day 10 for cohort 1 or all other cohorts, respectively) and remained resident in the CPU until 48 hours after the final dose (day 19 or day 26 for cohort 1 or all other cohorts, respectively). Thereafter, they returned to the CPU daily for 7 or 10 days for cohort 1 or all other cohorts, respectively, for safety monitoring and in cohort 5 for sample collection for bone turnover markers (BTMs). All subjects returned to the CPU 10 to 14 days after the final dose for a final visit, including sample collection for BTMs in cohorts 4 and 5.

Dosing

Subjects were dosed between 8.30 and 9.30 a.m. after an overnight fast and remained fasted for 2 hours after the dose to allow for any confounding effect of feeding on pharmacokinetic assessments, which were being performed as part of this phase I trial. Each subject received his dose at the same time on each occasion. Each received a single oral dose of saracatinib difumarate formulated as tablets or a matched placebo, in which saracatinib was substituted with an equal quantity of mannitol. The tablets were given with 240 mL of purified water while the subjects rested in a semirecumbent position.

All subjects were required to abstain from alcohol and caffeine-containing drinks or foods, which may affect physiologic measurements made during phase I assessments, and certain other foods, including grapefruit, from 48 hours before admission to the CPU until the end of each treatment visit. Saracatinib is metabolized mainly via CYP3A4, which is inhibited by grapefruit. They also were required to abstain from strenuous physical activity that was not within the subject's normal weekly routine during the trial. No concomitant medication or therapy except paracetamol was allowed during the study.

Safety monitoring

Heart rate, blood pressure, and 12-lead electrocardiograph were monitored throughout the study. Clinical chemistry, hematology, and urinalysis also were performed regularly throughout the study. Bleeding time and platelet aggregation were assessed immediately before the initial and final doses and at the end-of-study visit. Platelet aggregation also was measured at t_max, when t_max had been determined. Fecal occult blood was monitored during initial and multiple dosing phases until 48 hours after the final dose. Adverse-event data were collected throughout the study.

Sample collection for measurement of bone turnover markers

All blood samples and second-morning-void urine samples for BTM measurements were collected after an overnight fast. At all time points, samples from individual subjects were collected at the same time in the morning. During the course of the study, it was decided to investigate the offset of effect of saracatinib. Therefore, in cohort 4, additional samples were collected at the final visit. In cohort 5, additional samples were collected at 120 and 168 hours after the final dose and at the final visit.

Measurement of bone turnover markers

Serum samples were stored at –80°C and urine samples at –20°C until analysis. sCTX, uNTX/Cr, and serum tartrate–resistant acid phosphatase 5b (TRAcP-5b) were measured as markers of bone resorption, and serum procollagen propeptide of type I collagen (PINP) and serum bone-specific alkaline phosphatase (bone ALP) were measured as markers of bone formation. All samples collected from each individual subject were analyzed in the same batch.

sCTX was measured by electrochemiluminescent immunoassay (Roche Elecsys, Roche Diagnostics, Germany). sCTX was detectable in all samples. The interassay coefficient of variation (CV) was 4.7%. uNTX was measured by an enhanced chemiluminescent immunoassay (Vitros ECi, OrthoDiagnostics). The interassay CV was 7.9%. uNTX was expressed as a ratio to urinary creatinine (uCr), which was measured by dry-slide chemistry (Vitros 250, OrthoDiagnostics). The interassay CV was 2.2%. Serum TRAcP-5b was measured in duplicate by manual immunoassay (Bone TRAP, Suomen Bioanalytiikka Oy). The intraassay CV was 3.5%, and the interassay CV was 11%.

PINP was measured by electrochemiluminescent immunoassay (Roche Elecsys, Roche Diagnostics). The interassay CV was 9.1%. Bone ALP was measured by paramagnetic particle immunoassay (Access 2, Beckman Coulter). The interassay CV was 7.5%.

Ethics

The trial was approved by the Independent Ethics Committee, Quorn Research Review Committee, Old Dalby, Leicestershire LE14 3LB, UK, and was managed in a manner consistent with good clinical practice. All subjects gave informed written consent.

Role of the Funding Source

The protocol was devised by AstraZeneca in collaboration with RAH and RE and the trial performed by AstraZeneca. The measurement of markers of bone turnover was performed independently at the University of Sheffield, Sheffield, UK. All authors had full access to the data. The statistical analysis was performed and verified collaboratively by the University of Sheffield (RAH) and AstraZeneca (MR). AstraZeneca did not have the right to approve or disapprove publication of the completed manuscript.

Sample size and statistical analysis

The main trial was designed primarily as a safety and tolerability study. The number of subjects was based on the need to gain adequate safety, tolerability, and pharmacokinetic data while exposing as few subjects as possible to the study procedures. Therefore, no formal power calculations were performed.

The BTM data were not normally distributed; therefore, all analyses were performed using log-transformed data. Data from placebo-treated subjects in each cohort were pooled. The changes from baseline in BTMs at each time point were calculated as a baseline-scaled ratio (BSR), where BSR was the exponential of the difference between log of the value at the specified time point and at baseline.

ANCOVA models were used to analyze log BSR at 24 hours after the single dose, 24 hours after the final multiple dose, and at the follow-up visit using log baseline as a covariate and dose as a fixed factor. Exponentiated least-squares means from these models provided estimates of the geometric mean BSR at each dose level adjusted for differences in baseline between the dose groups. BSRs were converted to percentage change from baseline as 100 × (BSR – 1). Changes in the saracatinib-treated groups were compared with the change in the pooled placebo-treated subjects. To preserve the type I error at 5% for each BTM, a step-down procedure starting with the highest dose was used. Once a nonsignificant p value was obtained, no further comparisons were made. Jonckheere's method was used to determine whether the changes in BTMs were dose-dependent. The alternative hypothesis under this test is that the change for placebo ≤60 mg dose ≤125 mg dose ≤185 mg dose ≤250 mg dose, with at least one of these inequalities being strict. Analyses were performed using SAS Version 8.01 (SAS Institute, Inc., Chicago, IL, USA) and Statgraphics Plus Version 5.0 (StatPoint, Inc.).

Results

The first subject was enrolled in the study on September 22, 2003, and the study was completed on July 26, 2004. A total of 59 subjects were enrolled (mean age 34.6 years). Cohort 4 contained 11 subjects owing to a shortfall in recruitment. Demographics and BTM levels for the subjects are shown in Table 1. Cohorts and treatment groups were well balanced with respect to demographic characteristics and baseline levels of BTMs.

Table 1. Demographics and Baseline Characteristics, Mean (SD)

Baseline characteristics	Saracatinib treated					Placebo
	Cohort 1, 60 mg (n = 9)	Cohort 2, 125 mg (n = 9)	Cohort 3, 250 mg (n = 9)	Cohort 4, 250 mg (n = 8)	Cohort 5, 185 mg (n = 9)	Pooled (n = 5)
Age, years	32.6 (10.0)	34.6 (7.6)	35.0 (13.8)	37.4 (7.2)	28.7 (9.7)	37.5 (7.8)
Height, cm	178.3 (7.9)	177.2 (8.2)	175.2 (4.7)	177.6 (6.4)	183.6 (5.5)	178.6 (7.2)
Weight, kg	80.3 (11.6)	77.9 (12.6)	77.6 (10.2)	80.2 (9.1)	85.3 (9.1)	85.0 (10.2)
Body mass index, kg/m^2	25.2 (2.3)	24.7 (2.6)	25.3 (3.5)	25.3 (1.6)	25.3 (2.6)	26.6 (2.0)
BTM
sCTX, ng/mL	0.588 (0.274)	0.644 (0.379)	0.542 (0.288)	0.578 (0.217)	0.701 (0.418)	0.496 (0.186)
uNTX/Cr, nmolBCE/mmol	34.8 (16.7)	41.4 (33.9)	46.3 (34.8)	33.0 (7.1)	43.6 (32.1)	30.6 (11.3)
TRAcP-5b, U/L	3.02 (0.38)	3.42 (0.91)	3.35 (0.68)	3.12 (0.78)	4.06 (1.57)	3.13 (0.52)
PINP, ng/mL	59.2 (26.0)	57.3 (33.5)	62.0 (37.9)	55.1 (21.0)	80.2 (57.0)	50.4 (18.5)
Bone ALP, ng/mL	15.2 (3.6)	13.7 (4.6)	12.6 (3.8)	11.4 (1.4)	14.1 (5.0)	11.7 (3.8)

Forty-two subjects completed the trial (Fig. 2). Six subjects were discontinued owing to adverse events: Two subjects in cohort 1 (60 mg), one treated with placebo, had palpitations during the washout phase after receiving the initial single dose and one treated with saracatinib had anorexia, fatigue, and abnormal liver function tests on day 9 of the multiple-dose phase; one subject in cohort 2 (125 mg), who was treated with placebo, had an upper respiratory tract infection on day 12 of the multiple-dose phase; and three subjects in cohort 3 (250 mg) treated with saracatinib had noncardiac chest pain and chest tightness on days 6, 9, and 10, respectively, of the multiple-dose phase. Pending the evaluation of these adverse events, the remaining nine subjects in cohort 3 were discontinued on day 10 of the multiple-dose phase and the cohort terminated. The adverse events subsequently proved to be neither serious nor dose-limiting, and the 250 mg dose therefore was repeated in a new cohort, cohort 4. In addition to the nine subjects discontinued in cohort 3, two further subjects were discontinued in cohort 4 (250 mg): One saracatinib-treated subject discontinued because bleeding time specified in the study protocol was extended by more than 30 minutes on day 13 of the multiple-dose phase, and one placebo-treated subject was unwilling to continue after receiving the initial single dose.

Figure 2. Flow diagram of trial. ^aOne placebo-treated subject was withdrawn owing to an adverse event during the washout period following the single dose. ^bOne placebo-treated subject was unwilling to continue after the single dose. ^cOne saracatinib-treated subject was withdrawn owing to an adverse event after eight doses of the multiple-dose phase. ^dOne placebo-treated subject was withdrawn owing to an adverse event after 12 doses of the multiple-dose phase but was included in the analysis because samples were available 24 hours after the final dose. ^eCohort discontinued. ^fOne saracatinib subject withdrawn owing to study-specific criteria after 13 doses but included in the analysis because samples were available 24 hours after the final dose.

Doses of 60, 125, and 185 mg were well tolerated. Dose escalation was stopped before dose-limiting toxicity was reached. The maximum tolerated dose (MTD) was defined as 250 mg because it was thought that the next dose would be intolerable. Cohort 5, therefore, received 185 mg saracatinib. All subjects in cohort 5 completed the study. The placebo-treated subject in cohort 2 and the saracatinib-treated subject in cohort 4 who were discontinued during the multiple-dose phase had valid BTM measurements 24 hours after their final dose and therefore were included in the intention-to-treat analysis.

Response of bone turnover markers

The percentage changes in BTMs over the course of the study are shown in Figures 3 and 4. There were no significant changes compared with placebo in any marker 24 hours after the initial single dose except for sCTX at the two highest doses; geometric mean decrease adjusted for baseline was 16% [95% confidence (CI) 11–21%] at 250 mg (p < .001) and 11% (95% CI 3–17%) at 185 mg (p < .05).

Figure 3. Geometric mean percentage change in markers of bone resorption. The dark-shaded area represents the period during the single-dose phase and washout period when samples were collected for BTMs. The “predose” sample of this phase was collected before the initial single dose was given. The light-shaded area represents the period during the multiple-dose phase and the follow-up period when samples were collected for BTMs. The “predose” sample was collected before the final dose was given. ● = placebo; ▪ = 60 mg; ▴ = 125 mg; ♦ = 185 mg; and ▾ = 250 mg saracatinib, respectively. Error bars indicate SEM.

Figure 4. Geometric mean percentage change in markers of bone formation. The dark-shaded area represents the period during the single-dose phase and washout period when samples were collected for BTMs. The “predose” sample of this phase was collected before the initial single dose was given. The light-shaded area represents the period during the multiple-dose phase and the follow-up period when samples were collected for BTMs. The “predose” sample was collected before the final dose was given. ● = placebo; ▪ = 60 mg; ▴ = 125 mg; ♦ = 185 mg; and ▾ = 250 mg saracatinib, respectively. Error bars indicate SEM.

Table 2 shows the percentage changes, adjusted for baseline values, 24 hours after the final dose of the multiple-dose phase. There were significant dose-dependent decreases from baseline in sCTX at all doses of saracatinib when compared with placebo (one-sided Jonckheere p < .0001). There were somewhat smaller dose-dependent decreases in uNTX/Cr (one-sided Jonckheere p < .05) that were significant when compared with placebo at doses greater than 60 mg. At 250 mg (MTD), there was an 88% (95% CI 84–91%) decrease from baseline in sCTX and a 67% (95% CI 53–77%) decrease in uNTX/Cr. At all doses greater than 60 mg, there was a significant decrease compared with placebo in TRAcP-5b 24 hours after the final dose. The size of the decrease ranged from 11% to 14%. There were no significant changes in the bone-formation markers PINP and bone ALP in response to the highest dose of saracatinib (250 mg) compared with placebo (see Table 2); therefore, lower doses were not compared, as noted under statistical methods.

Table 2. Percentage Change in Markers from Baseline at 24 Hours After Final Dose, Geometric Mean Adjusted for Baseline (95% CI)

BTM	Placebo (n = 10)	60 mg (n = 8)	125 mg (n = 9)	185 mg (n = 9)	250 mg (n = 8)
^nsNot significant; ^*p < .05; ^**p < .01; ^***p < 0.001; compared with placebo. ^naNot assessed. For each marker, percentage change at 24 hours after the final dose was compared with placebo in a dose-descending order until the first nonsignificant p value was found, after which no further comparisons were made.
sCTX	17 (–7, 49)	–24^* (–41, –1)	–55^*** (–65, –43)	–71^*** (–77, –63)	–88^*** (–91, –84)
uNTX/Cr	5 (–23, 44)	–3^ns (–32, 37)	–39^* (–56, –15)	–69^*** (–78, –56)	–67^*** (–77, –53)
TRAcP-5b	9 (0, 17)	2^ns (–7, 11)	–13^*** (–19, –5)	–14^*** (–21, –6)	–11^** (–18, –3)
PINP	17 (–1, 38)	3^na (–14, 24)	23^na (3, 46)	33^na (11, 58)	13^ns (–6, 35)
Bone ALP	3 (–5, 11)	5^na (–4, 15)	10^na (1, 19)	–3^na (–11, 5)	9^ns (0, 19)

To observe the offset of effect, additional measurements of bone markers were made after the cessation of treatment in the cohorts treated with the two highest doses of saracatinib. At the final visit at the end of the study, 10 to 14 days (mean 12 days) after the final dose, levels of sCTX and uNTX/Cr in both cohorts had returned toward baseline. However, the decrease from baseline at the poststudy visit in the subjects treated with saracatinib still remained significantly greater than in the placebo-treated subjects (Table 3). In contrast, TRAcP-5b continued to decrease after the cessation of treatment. Although cessation of the 250 mg dose of saracatinib resulted in a significant decrease in PINP (28%; 95% CI 13–40%), in general, offset of treatment had no effect on markers of bone formation.

Table 3. Percentage Change in Markers from Baseline at Final Visit (Mean 12 Days After Final Dose), Geometric Mean Adjusted for Baseline (95% CI)

BTM	Placebo (n = 5)	185 mg (n = 9)	250 mg (n = 8)
^nsNot significant; ^*p < .05; ^**p < .01; ***p < .001; ^****p < .0001; compared with placebo. ^naNot assessed. For each marker, percentage change at the final visit was compared with placebo in a dose-descending order until the first nonsignificant p value was found, after which no further comparisons were made.
sCTX	–15 (–37, 16)	–56^** (–65, –44)	–67^**** (–74, –58)
uNTX/Cr	–5 (–40, 50)	–44^ns (–60, –21)	–49^* (–64, –26)
TRAcP-5b	–5 (–13, 4)	–25^*** (–30, –20)	–30^**** (–35, –25)
PINP	6 (–16, 33)	1^ns (–15, 21)	–28^* (–40, –13)
Bone ALP	–1 (–8, 8)	–7^na (–13, –1)	–5^ns (–11, 1)

Adverse events

The most commonly reported adverse events were headache [21 saracatinib (39%) versus 4 placebo (26%)], papular rash [16 saracatinib (30%) versus 1 placebo (6%)], upper respiratory tract infection [10 saracatinib (19%) versus 4 placebo (27%)], loose stools [13 saracatinib (24%) versus 0 placebo (0%)], and myalgia [10 saracatinib (19%) versus 1 placebo (6%)].

Protocol deviations

Cohort 3 was discontinued owing to suspected serious adverse events, as described earlier. There were no other significant protocol deviations.

Discussion

In vitro and animal studies have shown that inhibition of Src family kinases significantly inhibits osteoclastic bone resorption.10, 13, 14 However, to date, the effect of Src inhibitors on bone resorption in humans has not been reported. In this phase I study, we have demonstrated that one such inhibitor, saracatinib, significantly reduces bone resorption in healthy men. We observed large decreases of more than 67% in type I collagen degradation products, sCTX and uNTX/Cr, in response to daily dosing with 185 or 250 mg of saracatinib over a 14 day period followed by a gradual return toward pretreatment levels after the cessation of treatment, although both markers were still significantly reduced compared with baseline 12 days after the final dose of saracatinib. This decrease in bone resorption could result from a reduction in osteoclast activity, a reduction in osteoclast numbers, or a combination of both. In the Src^–/– mouse, there are failures in osteoclast cytoskeleton and podosome formation 15 and in ruffled border formation 6 that lead to the inability to resorb bone. Mullender and colleagues 10 have shown that saracatinib inhibits bone resorption by reducing both the number and size of the bone pits in an in vitro study using rabbit osteoclasts cultured on bovine bone slices. These data suggest that at least some of the reduction in bone resorption we observed was due to a decrease in osteoclast activity.

In addition to measuring sCTX and uNTX/Cr, we also measured TRAcP 5b, which is generally regarded as a marker of osteoclast number rather than activity.16 We observed a small decrease in TRAcP 5b, suggesting some decrease in osteoclast number. A reduction in osteoclast number could result from a decrease in the number of osteoclasts reaching maturity or an increase in the number undergoing apoptosis. A recent in vitro study using peripheral blood mononuclear cells has shown that saracatinib may inhibit osteoclast differentiation,10 and Recchia and colleagues17 have shown that another Src inhibitor, CGP76030, induces apoptosis in osteoclasts in bone marrow cultures, supporting both these possibilities.

In the Src^–/– mouse, the number of mature osteoclasts on the bone surface is the same or increased as compared with the wild-type mouse.5, 6 The apparent discrepancy between this and our findings, which suggest that the effect of saracatinib is mediated, in part, by a reduction in the number of osteoclasts, can be explained by saracatinib being an inhibitor of all Src family kinases. The IC₅₀ values for other members of the Src family such as cYes, Lck, and Lyn are 4 nM, <4 nM, and 5 nM, respectively, compared with 2.7 nM for cSrc.8 It is therefore possible that given a large amount of functional redundancy in the Src family kinases, other members of the Src family substitute for Src in the Src^–/– mouse, whereas in our study other Src family members were unable to take on the role of Src because they too were inhibited. Our findings support an effect of saracatinib to both inhibit osteoclast activity and to reduce osteoclast number.

Interestingly, in contrast to CTX and NTX/Cr, in the two cohorts treated with the highest doses, TRAcP-5b continued to decrease after treatment was stopped. This divergent response of telopeptides (CTX and NTX/Cr) and TRAcP-5b to offset of treatment suggests that there may be temporal differences in recovery after Src family kinase inhibition of the inactivated osteoclasts and the restoration of osteoclastogenesis or suppression of apoptosis. Alternatively, it may be that the half-life of circulating TRAcP-5b is longer than that of the telopeptides, and hence the response to change in treatment appears to be slower.

We measured bone ALP and PINP as markers of bone formation but did not observe any significant changes in response to saracatinib. Bone resorption and bone formation are normally “coupled” events; consequently, treatment with antiresorptive agents such as bisphosphonates leads to a decrease in bone resorption followed, several weeks later, by a decrease in bone formation.18 In this study, the subjects were treated for only 14 days, which probably was too short to see any significant effect on bone formation. However, there may be other explanations. A new group of drugs is emerging, including the chloride channel inhibitors,19 that appears to “uncouple” bone resorption and formation, suppressing bone resorption while bone formation remains unaffected. This may be due to the ability of nonresorbing osteoclasts to stimulate osteoblastic bone formation 20 via non-bone-derived factors. However, in the case of Src inhibition, there also may be a direct effect on bone formation. Marzia and colleagues 7 have demonstrated that osteoblast differentiation and bone formation are increased in primary and immortalized osteoblasts from Src^–/– mice and osteoblasts from normal mice treated with an Src antisense oligodeoxynucleotide sequence. Our failure to observe any actual increase in bone formation may be due to differences in dose required to inhibit osteoclasts and stimulate osteoblasts or to the length of the treatment, or it may have been counteracted by the effect of saracatinib on Abl kinase because Abl kinase deficiency has been shown to lead to defects in osteoblast maturation and mineral deposition.9

The proportion of adverse events reported by subjects in the active arms of this study was greater than in subjects in the placebo arm. This is not unexpected because one aspect of the design of this phase I trial was to assess the tolerability and define the MTD of saracatinib.

The decrease in bone resorption observed, even with the 185 mg dose, a dose below the MTD, is comparable with the decrease observed in postmenopausal women treated with bisphosphonates.21 Reductions of bone turnover of this magnitude are associated with increased bone mass and reduction of fracture risk.22 Our findings therefore support and add weight to earlier proposals,23 based on in vitro and animal studies, that Src kinase inhibitors are potential candidate drugs in the treatment of osteoporosis and other metabolic bone diseases associated with increased bone resorption, in particular metastatic bone disease. Src kinase inhibitors have been shown, in animal studies, to reduce the incidence of bone metastases.14 This may be due to a dual effect of the inhibitor on the differentiation, survival, motility, and angiogenesis 24 of the malignant cells themselves and on osteoclastic bone resorption.

In conclusion, we have shown that an Src family kinase inhibitor, saracatinib, can significantly reduce bone resorption in healthy men at doses that were well tolerated. Bone formation was not substantially affected by the treatment, and therefore, longer studies are required to fully investigate the effect of Src inhibition on bone formation. Saracatinib is a potential therapeutic agent for the treatment of metabolic bone diseases associated with elevated bone resorption, such as metastatic bone disease and osteoporosis. Further trials are ongoing in the patients with solid tumors with and without metastatic bone disease.

Disclosures

RAH has received research funding from AstraZeneca and Procter & Gamble Pharmaceuticals and honoraria from Pfizer. RE has received consultancy fees, honoraria, and research grants from Amgen, AstraZeneca, California Pacific Medical Center, GlaxoSmithKline, Hologic, Kyphon, Inc., Lilly Industries, Maxygen, Nastech Pharmaceuticals, Nestle Research Center, New Zealand Milk Limited, Novartis, Novo Nordisk, ONO-Pharma, Organon Laboratories, Osteologix, Pfizer, Procter & Gamble Pharmaceuticals, Roche Diagnostics, Sanofi-Aventis, Servier, Shire, Thethys, Transpharma Medical Limited, Unilever and Unipath. GC, AS, and RDF are employees of AstraZeneca and own stock in AstraZeneca. At the time of the study and manuscript preparation, MR and JAL were employees of AstraZeneca and owned stock in AstraZeneca.

RAH, GC, AS, JAL, and RE contributed to the concept and design of the study; JAL contributed study material; RAH, GC, MR, and JAL contributed to data collection; and RAH, GC, MR, and RE contributed to manuscript preparation. All authors contributed to data analysis and interpretation, critically reviewed the manuscript, and provided final approval of the version to be published.

Acknowledgements

We wish to acknowledge Matt Lewis, PhD, of Mudskipper Bioscience for editorial comments and preparation of figures. This study was funded by AstraZeneca.