A Phase I Study of the HSP90 Inhibitor Retaspimycin Hydrochloride (IPI-504) in Patients with Gastrointestinal Stromal Tumors or Soft-Tissue Sarcomas

Sarcomas are a heterogeneous group of malignant tumors linked by a common mesenchymal origin (1). Gastrointestinal stromal tumors (GIST), the most common subset of soft-tissue sarcomas (STS), harbor driver gene mutations in approximately 80% of cases involving KIT or, in a separate 15%, PDGFRA; these mutations result in ligand-independent signaling through the activated KIT or platelet-derived growth factor receptor α (PDGFRα) receptor tyrosine kinases (2, 3). The remaining GISTs do not have identified tyrosine kinase mutations but often have genomic aberrancies in succinate dehydrogenase subunits (4). Small-molecule tyrosine kinase inhibitors (TKI), such as imatinib and sunitinib, block KIT and PDGFRα signaling, resulting in significantly improved outcomes for patients with advanced GIST as well as in patients who receive adjuvant therapy following resection of primary localized GIST at high risk of recurrence (5–10). Despite the dramatic improvements in the treatment of advanced GIST, the majority of patients who initially benefit from TKIs will eventually develop polyclonal GIST that is resistant to TKI therapy, most often due to emergence of secondary KIT mutations (6, 11–16). Novel strategies that bypass specific mutations which result in structural resistance to TKIs in order to target fundamental pathways that support aberrant oncoprotein activation might be effective in TKI-resistant GIST.

Heat shock protein 90 (HSP90) is a molecular chaperone that uses the energy of ATP hydrolysis to increase the stability and activity of its client proteins, many of which are involved in key pathways important in malignancy (17–19). The repertoire of client proteins of HSP90 includes a host of proteins important to tumor growth and proliferation. These proteins influence the hallmark traits of cancer such as growth factor independence, resistance to antigrowth signals, unlimited replicative potential, tissue invasion and metastasis, evasion of apoptosis, and sustained angiogenesis (20). The HSP90 chaperone complex facilitates the conformational maturation, stability, and activation of numerous wild-type and mutated oncoproteins, including KIT and PDGFRα while preventing proteasome-mediated degradation (21). Because activating mutations in KIT and PDGFRα play a critical role in the pathogenesis and progression of GIST, and because the cellular proteins from these genes rely on HSP90 for optimal processing and function, we hypothesized that HSP90 inhibition could have beneficial effects in patients with GIST. Preclinical models supported this hypothesis (11), and this clinical trial was the translation of that concept to patients.

IPI-504 is an HSP90 inhibitor uniquely designed to overcome the therapeutic limitations of earlier HSP90 inhibitors with potential advantages over those currently in development (22, 23). IPI-504 is the hydrochloride (HCl) salt of the hydroquinone of 17-AAG (a quinone) and in vivo, the hydroquinone and quinone exist in a dynamic equilibrium (23). Inside cells, 17-AAG is enzymatically reduced to the hydroquinone (free base of IPI-504) which is a 40- to 60-fold more potent inhibitor of HSP90 than 17-AAG (24, 25). IPI-504 is also subject to oxidative metabolism via CYP3A4 to produce the active metabolite 17-amino-17-demethoxygeldanamycin (17-AG). Previous clinical studies of 17-AAG showed some evidence of activity but were limited by formulation concerns and toxicities due to these formulations [e.g., Cremophor, dimethyl sulfoxide (DMSO)] that were required because of aqueous insolubility (23, 26–28). In human GIST cell lines characterized either by sensitivity or resistance to imatinib, HSP90 inhibitors, including IPI-504, suppress cell growth and inhibit activation of oncogenic KIT and PDGFRα as well as the relevant downstream signaling molecules [e.g., AKT and mitogen-activated protein kinase (MAPK); refs. 11, 29, 30]. In addition, single-agent IPI-504 treatment significantly reduces tumor volume in GIST xenografts (31). Other STS, driven by fusion proteins that are the result of chromosomal translocations, such as Ewing sarcoma and synovial sarcoma, may also be sensitive to HSP90 inhibition (32, 33). On the basis of these promising preclinical data, a phase I trial was conducted to determine the safety and the recommended phase II dose and schedule of IPI-504 in patients with advanced (metastatic and/or unresectable) imatinib-resistant GIST or other STS.

Patients were eligible if they had histologic confirmation of either (i) metastatic and/or unresectable GIST that was imatinib-resistant or (ii) another form of STS not amenable to curative therapy by conventional multimodality options. Main inclusion criteria were as follows: ≥18 years old; primary or secondary resistance to, or unacceptably severe medical intolerance of, imatinib for patients with GIST; and an Eastern Cooperative Oncology Group (ECOG) performance status of 0–2. Patients were excluded for any of the following: prior exposure to any known HSP90 inhibitor; treatment with an investigational drug or kinase inhibitor within 2 weeks of study start; NYHA class 3 or 4 congestive heart failure or left ventricular ejection fraction <40%; prior radiation therapy that potentially included the heart in the field; myocardial infarction or active ischemic heart disease within 6 months; history of arrhythmia; baseline QTc ≥ 450 msec; congenital long QT syndrome or first-degree relative with unexplained sudden death under 40 years of age; left bundle branch block on electrocardiogram (ECG); grade ≥ 3 peripheral neuropathy; serum creatinine > 1.5 × upper limit of normal (ULN); alkaline phosphatase > 2.5 × ULN; amylase and lipase > 1.5 × ULN; aspartate aminotransferase (AST) or alanine aminotransferase (ALT) > 2.5 × ULN; hemoglobin <9.0 g/dL; platelets <100,000 mm³; absolute neutrophil count <1,500 cells/mm³; active infection or systemic use of antibiotics within 72 hours of treatment; or breast feeding. Concurrent treatment with radiation therapy, agents that alter CYP3A or agents that may prolong the QTc interval were not permitted. All patients provided written informed consent, and the study was approved by the local institutional review boards (Trial Registration ID: NCT00276302).

This was a multicenter, phase I, open-label, sequential intercohort dose-escalation study designed primarily to evaluate safety, identify the MTD, and determine the recommended phase II dose and schedule of IPI-504 for patients with metastatic and/or unresectable GIST or other STS. Secondary objectives were to determine pharmacokinetic parameters, assess antitumor activity, and explore biologic markers of IPI-504 activity. The study sponsor was Infinity Pharmaceuticals, Inc., which also provided IPI-504 (sterile lyophilized powder).

IPI-504 was administered as a 30-minute intravenous infusion via 1 of 2 different schedules: schedule 1 comprised twice weekly dosing (days 1, 4, 8, and 11) for 14 consecutive days followed by a 10-day rest period, whereas schedule 2 administered twice-weekly dosing (days 1, 4, 8, 11, 15, and 18) for 21 consecutive days. Safety and efficacy data are reported here only from schedule 1, due to early termination of schedule 2 (data from schedule 2 have been presented elsewhere; ref. 34). However, pharmacokinetic parameters were determined and reported here from patients on both schedules. Although study drug administration was generally stopped if disease progression occurred, the study allowed continued study drug administration if patients were deriving clinical benefit in the judgment of the local investigator.

On the basis of preclinical toxicology data, the starting dose of IPI-504 was 90 mg/m². Dose escalation in subsequent cohorts followed the modified Fibonacci 3 + 3 design, increasing in increments of approximately 66%, 50%, 33%, 33%, 25%, and 25% of the preceding dose level in the absence of dose-limiting toxicities (DLT). If 1 of the 3 initial patients experienced a DLT, the cohort was expanded to 6 patients. The MTD was defined as the highest dose at which fewer than 33% of at least 6 patients experienced a DLT during the first cycle of treatment during the dose escalation phase. Twenty-six additional patients were enrolled at the MTD to evaluate safety with greater confidence.

DLTs included any of the following events that occurred during the first cycle and were attributed as at least possibly related to study drug: death, grade 4 neutropenia lasting ≥5 days, grade 3 febrile neutropenia with elevated temperature (>101.5°F) confirmed on 2 occasions, grade 4 thrombocytopenia with platelets <10,000/mm³, any other hematologic toxicity of grade ≥4, grade 3 nausea, vomiting, or diarrhea despite maximal medical management, QTc prolongation [grade 3 if it occurred on day 1 at 5 minutes or 1 hour post initial infusion; grade 2 if it occurred on day 1 at 6 hours post initial infusion, or at any subsequent baseline ECG, or if it occurred for the first time during a subsequent infusion (days 4, 8, and 11) at baseline or 5 minutes], and any other nonhematologic toxicity of grade ≥3 with the exception of asymptomatic lipase elevation in the absence of any clinical signs or symptoms of pancreatitis.

Patients underwent clinical safety assessments including physical examination, history, concomitant medications, ECGs, clinical laboratory tests and adverse events. Adverse events were graded according to the National Cancer Institute Common Terminology Criteria for Adverse Events (NCI CTCAE) version 3.0. ECGs were obtained in triplicate within a 5-minute period. During cycle 1, ECGs were conducted on day 1 at baseline and 5, 60, and 360 minutes post dose administration, on days 4, 8, and 11 pre-dose and 5 minutes post-dose administration, and on day 21 pre-dose only.

Blood samples for pharmacokinetic analyses were obtained at baseline on cycle 1, day 1 before IPI-504 infusion and serially at 5 minutes, 30 minutes, and 1, 2, 6, and 24 hours after the end of the infusion. Additional samples were collected on cycle 1, day 11 (dose 4) before infusion and 5 minutes, 30 minutes, and 1 hour after the end of the infusion. Pharmacokinetic parameters were determined from 60 patients following administration of the dose on cycle 1, day 1.

Plasma concentrations of IPI-504, 17-AAG, and 17-AG were determined using a validated liquid chromatographic method with mass spectrometry detection. The following parameters were determined for IPI-504, 17-AAG, and 17-AG on cycle 1 day 1 using noncompartmental pharmacokinetic analysis methods: maximum concentration (C_max), the time of maximum concentration (T_max), area under the concentration–time curve extrapolated to infinity [AUC_(0–∞)], and terminal elimination half-life (t_½). In addition, systemic clearance (CL) and volume of distribution at steady state (V_ss) were determined for IPI-504. Pharmacokinetic parameters were calculated for each patient using the actual time of sample collection relative to the beginning of infusion.

Peripheral blood leukocytes (PBL) were isolated from blood collected before and 24 hours following the first IPI-504 dose of cycle 1. PBL protein lysates were subjected to SDS-PAGE, transferred to nitrocellulose membranes, and incubated first with primary anti-Hsp70 antibodies (Cell Signaling Technology, Inc.) and then with secondary antibodies (goat anti-mouse HRP antibodies; Amersham Biosciences). Proteins were visualized by chemiluminescence (ECL, Amersham Biosciences) on BioMax film.

Tumor response was assessed per RECIST v1.0 (35). Computed tomographic (CT) scans were obtained at baseline and at the end of cycle 1. After cycle 1, CT scans were obtained approximately every 2 cycles or according to institutional standard.

Functional imaging with 18-FDG-PET using maximal standardized uptake values (SUV_max) was conducted optionally to gain additional insight to potential antitumor activity of IPI-504 in patients with FDG-avid disease at baseline. PET scans were obtained at baseline, after the second or third dose of IPI-504 in cycle 1, and again at the end of the “drug holiday” in the first cycle. PET responses were evaluated according to European Organization for Research and Treatment of Cancer (EORTC) criteria, using SUV_max values as quantification metric (36).

Progression-free survival (PFS) was measured from the date of first dose to the date of documented progression or death, whichever came first. Time to tumor progression (TTP) was the duration in weeks from the date of the first dose until the date of confirmed progressive disease (PD) using conventional RECIST definitions. Survival analyses were carried out using Kaplan–Meier life test methods.

Between December 19, 2005 and June 7, 2010, a total of 54 patients (37 GIST, 17 STS) were enrolled on study and received IPI-504 twice weekly for 2 weeks on/1 week off in 6 dosing cohorts, ranging from 90 to 500 mg/m². Baseline characteristics and disease history are summarized in Table 1. All patients with GIST had received prior therapy with imatinib, and the majority (78%) received an imatinib dose greater than 400 mg after initial progression. In addition, more than one third (38%) of patients with GIST also received prior sunitinib, and the majority (81%) of patients had received 3 or more prior therapies. Patients with STS had received a median of 6 prior anti-cancer therapies.

One patient in the first dose level (IPI-504 90 mg/m²) experienced asymptomatic elevation in circulating lipase, which was initially considered a DLT before a protocol amendment that excluded asymptomatic lipase elevation from the DLT definition in the absence of any clinical signs or symptoms of pancreatitis. No other DLTs were observed in the first 4 dose levels (90–300 mg/m²). A single DLT of grade 5 intracerebral hemorrhage occurred in the 400 mg/m² dose level. This cohort was expanded to 6 patients, with no further DLTs observed at this dose level. Of the 6 patients in the 500 mg/m² cohort, a DLT of grade 3 headache occurred in 1 patient, and a second DLT of grade 3 myalgia occurred in another patient. Because the rate of DLTs in the 500 mg/m² cohort exceeded the limit defined in the protocol, the MTD was established as 400 mg/m² administered twice weekly for 2 weeks on/1 week off, and an additional 26 patients were enrolled at this dose and schedule for further evaluation.

Patients were on study for a median of 53 days (range, 4–824) and received a median of 11 doses of IPI-504 on study; one patient with GIST who continues to receive IPI-504 at the time of this publication through a single-patient access program has been on study for approximately 5 years. Dose reductions were required in 7 of 32 patients in the 400 mg/m² cohort, and 3 of 7 patients in the 500 mg/m² cohort, most commonly for fatigue. Interruptions of study drug dosing occurred in 24 (44%) patients and were not dose-dependent. Reasons for discontinuation from the study included PD in 26 (48%) patients, other reasons in 20 (37%) patients (typically attributed to disease progression), adverse events in 3 (5.6%) patients (two cases of liver function test elevation and one of pain), withdrawal by subject in 3 (5.6%) patients, and death in 2 (3.7%) patients.

All patients reported at least one treatment-emergent adverse event during the study. The most common adverse events regardless of causality were fatigue, nausea, headache, diarrhea, and vomiting (Table 2), with the majority being grade 1 or 2. Grade 1 sinus bradycardia and tachycardia occurred in 11 (20%) and 5 (9%) patients, respectively. Grade 1 and grade 2 QTc prolongation occurred in 9 (17%) and 1 (2%) of patients, respectively, and no patients had a QTc increase > 60 msec or a QTc > 475 on study. A more detailed account of the cardiac monitoring of patients receiving IPI-504 has been presented previously (37).

The most common grade 3 and 4 related adverse event with a frequency ≥5% were fatigue (9.3%), increased AST (7.5%), and increased ALT (5.6%).

Two patients died on study of causes considered at least possibly related to therapy. A 68-year-old male patient with GIST and numerous metastases to the lung, liver, peritoneum, thoracic vertebrae, and left acetabulum developed hepatic and renal failure leading to death after receiving two doses of 400 mg/m² IPI-504. A 60-year-old female patient with GIST and one metastasis to the liver had intracerebral hemorrhage leading to death after receiving 2 doses of 400 mg/m² IPI-504. Autopsy did not reveal structural causes for this hemorrhage.

Serious adverse events were reported for 20 (37%) patients. Seven patients experienced serious adverse event that were considered at least possibly related to therapy, including dehydration, diarrhea, nausea, vomiting, increased AST, fatigue, headache, increased lipase, myalgia, and the 2 grade 5 events described above.

Following administration of the first dose of IPI-504, the mean concentration profiles of IPI-504 and 17-AAG displayed similar shapes, with peak plasma concentrations observed at the end of the infusion, followed by bi-exponential decline (Supplementary Fig. S1). Peak plasma concentrations of 17-AG were observed slightly later, followed by mono-exponential decline. Pharmacokinetic parameters for IPI-504, 17-AAG, and 17-AG following IPI-504 administration on cycle 1 day 1 are summarized in Supplementary Table S1. At the MTD of 400 mg/m², C_max values for IPI-504 and 17-AAG were 6,740 and 7,823 ng/mL, respectively, whereas the C_max for 17—AG was slightly lower (3,381 ng/mL). On average, exposure to 17-AAG approximated that of IPI-504, whereas 17-AG exposure was typically twice that of the parent drug. Overall, exposure to 17-AG accounted for about 50% of the total exposure, determined as the sum of the AUCs for IPI—504, 17-AAG, and 17-AG.

Mean clearance (CL) values for IPI-504 across the evaluated doses ranged from 70.9 to 121 L/h, with the exception of the 300 mg/m² dose group, where CL was estimated to be 141 L/h. Volume of distribution (V_ss) values averaged 119 to 275 L across the dose levels, with the exception of the 300 mg/m² dose level (V_ss = 291). Mean elimination half-life values for IPI-504 and 17-AAG ranged from 1.4 to 3.3 hours after single dose administration; 17-AG elimination was slightly slower, with a mean elimination half-life of 4.3 to 7.5 hours. There was no accumulation of IPI-504 or its metabolites observed after repeated dosing on a twice weekly administration schedule based on a comparison of the AUCs observed through 1.5 hours at the end of cycle 1 to the first dose in cycle 1 (data not shown).

Induction of Hsp70, which was used as a marker for HSP90 inhibition, was observed in 8 of 10 patients from whom evaluable PBL samples were obtained. All 7 patients treated at the 225 mg/m² dose level or higher showed a pharmacodynamic response, and 1 of the 2 patients treated with 150 mg/m² also showed an increase in Hsp70 protein expression. The patient treated at 90 mg/m² did not show an induction of Hsp70 expression (Supplementary Fig. S2).

Antitumor clinical activity was assessed in all patients (n = 54). In the 37 patients with GIST, the overall objective response rate was 3% (1 patient with a PR), with 73% of patients achieving SD or PR for at least 6 weeks, and 16% at 12 weeks. RECIST outcome for patients with GIST is shown in Fig. 1A. Before study participation, the patient with a PR had had progression of disease on imatinib (400 and 800 mg) and sunitinib (37.5 and 50 mg) and has been receiving IPI-504 for approximately 5 years at the time of this publication (Fig. 1B).

KIT genotyping was available for 15 patients with GIST. Of these, 8 had primary mutations in exon 11, 4 in exon 9, 1 in exon 13, and 2 were without identified mutations in KIT, including the tumor that exhibited a PR. Six tumors had secondary mutations in exon 13 or 17. There was no correlation between outcome and primary mutation; however, 3 of 5 sequenced tumors with minor or partial responses had secondary mutations, compared with 3 of 10 sequenced tumors with disease progression as best response. While it is difficult to make any conclusions based on this small sample size, these data are consistent with preclinical data suggesting that more heavily mutated KIT is more sensitive to HSP90 inhibition than singly mutated KIT (11).

Of the 17 patients with STS, the overall response rate was 6% (1 patient with metastatic dedifferentiated liposarcoma exhibited a durable PR for 2.25 years after initiation of IPI-504), with 59% of patients having SD for at least 6 weeks, and 18% at 12 weeks (Table 3 and data not shown). RECIST outcome for patients with STS is shown in Fig. 1C.

18-FDG-PET analysis was conducted on 29 of 37 patients with GIST. Responses were evaluated according to the EORTC criteria (37). Of the 29 patients in whom PET imaging was conducted, the overall response rate was 38% (all metabolic PRs). Nineteen (66%) patients had some overall reduction in FDG avidity, whereas 2 (7%) patients exhibited metabolic progression of disease as best response (Fig. 2). PET imaging conducted during the drug holiday week of cycle 1 showed variable changes in SUV_max compared to on-drug imaging. All 4 patients with FDG-PET partial responses on C1D11 had a rebound of SUV_max on C1D21 but it remained less than baseline imaging, with 1 patient continuing with a metabolic PR. Of the 8 patients with minor reductions in SUV_max on C1D11, 2 had further reduction on C1D21 (1 resulting in a metabolic PR), 2 showed no change between C1D11 and C1D21, 1 showed less of a decrease, and 2 showed an increase in SUV_max (1 resulting in metabolic PD). Although a “flare phenomenon” was observed in some patients, there was no clear pattern in regard to changes in FDG-PET imaging during the washout week.

Median PFS was 9.1 weeks [95% confidence interval (CI), 6.7–12.0] for all patients and 10.6 weeks (95% CI, 6.4–12.1) for patients with GIST (Fig. 3). Median TTP was 10.6 weeks (95% CI, 8.9–12.1) for all patients and 12 weeks (95% CI, 9.0–12.6) for patients with GIST. There were no obvious correlations between changes in FDG-PET and either CT response or duration of disease control.

In this phase I open-label, dose-escalation study, IPI-504 was generally well tolerated by patients. Moreover, evidence of antitumor activity was observed, most frequently as metabolic responses by functional ¹⁸FDG-PET imaging in patients with GIST. In this heavily pretreated population, all patients with GIST had previously progressed on imatinib therapy, and many had also received prior sunitinib treatment. Nearly 40% of patients with GIST had received 3 or more therapies before study entry. Despite this degree of pretreatment, 38% of patients with GIST had a metabolic partial response to IPI-504. The disease control rate (defined as a RECIST PR, CR, or SD for at least 6 weeks) was 73% for the patients with GIST, despite the low overall objective response rate (3%, representing a single patient with GIST).

On the basis of these data, an international randomized phase III trial of IPI-504 at this dose and schedule (400 mg/m² twice weekly for 2 weeks on/1 week off) was initiated in patients with advanced GIST following failure of prior imatinib and sunitinib. This trial was terminated early, however, due to safety concerns of hepatic toxicity in the IPI-504 arm, which was both more common and more severe in the phase III trial, highlighting the challenges of moving from a small phase I experience to broader studies, even in the same patient population (38). Further clinical studies have shown that a once-weekly schedule of administration is associated with greater hepatic safety than is a twice-weekly schedule of administration, and clinical studies of IPI-504 moving forward are incorporating once-weekly dose administration.

No clinically meaningful effects on cardiac function or repolarization were observed in this study. While episodes of first-degree AV block, sinus bradycardia, and sinus tachycardia were observed occasionally, none of these adverse events ever exceeded grade 1, and all were rapidly reversible without requiring a dose modification or any other intervention. Importantly, the QTc interval was not significantly affected with IPI-504 treatment, consistent with previous findings indicating that there is no relationship between 17-AAG and the QT interval among patients with advanced cancers (39).

In conclusion, in this phase I study administered to patients with advanced GIST or other STS, intravenous IPI-504 was generally well tolerated with evidence of antitumor activity. Importantly, however, further clinical work has shown that IPI-504 at 400 mg/m² twice weekly for 2 weeks on/1 week off is not an appropriate dose and schedule in this patient population. IPI-504 is currently being evaluated in a phase II, randomized, double-blind, placebo-controlled study with a once-weekly schedule of administration in combination with docetaxel in patients with stage IIIb/IV non–small cell lung cancer (NSCLC) with a history of heavy smoking.

A.J. Wagner has commercial research grants from and is a Consultant/Advisory Board member of Infinity. R. Chugh and L.S. Rosen have commercial research grants from Infinity Pharmaceuticals. S. George is a Consultant/Advisory Board member of Ariad and Pfizer. G.D. Demetri has commercial research grants from Infinity, Synta, Bayer, Novartis, and Pfizer and is a Consultant/Advisory Board member of Infinity, Pfizer, Bayer, and Novartis. J. Dunbar, E. Normant, and D. Grayzel are current or former employees of Infinity Pharmaceuticals, Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: L.S. Rosen, M.S. Gordon, D. Grayzel, G.D. Demetri

Development of methodology: L.S. Rosen, M.S. Gordon, J. Dunbar, D. Grayzel, G.D. Demetri

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.J. Wagner, R. Chugh, L.S. Rosen, J.A. Morgan, S. George, M.S. Gordon, E. Normant, G.D. Demetri

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.J. Wagner, R. Chugh, L.S. Rosen, M.S. Gordon, J. Dunbar, E. Normant, D. Grayzel, G.D. Demetri

Writing, review, and/or revision of the manuscript: A.J. Wagner, R. Chugh, L.S. Rosen, J.A. Morgan, S. George, M.S. Gordon, J. Dunbar, D. Grayzel, G.D. Demetri

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.J. Wagner, G.D. Demetri

Study supervision: L.S. Rosen, G.D. Demetri

Dana-Farber/Harvard, University of Michigan, UCLA and Pinnacle Oncology received research funding from Infinity Pharmaceuticals, Inc. to support the conduct of this trial.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.