doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`TRIAL
`
`Evidence of CFTR Function in Cystic Fibrosis after
`Systemic Administration of 4-Phenylbutyrate
`Pamela L. Zeitlin,1,* Marie Diener-West,3 Ronald C. Rubenstein,5 Michael P. Boyle,2
`Carlton K. K. Lee,4 and Lois Brass-Ernst1
`
`Departments of 1Pediatrics and 2Medicine, Johns Hopkins University School of Medicine, 3Department of Biostatistics,
`Johns Hopkins Bloomberg School of Public Health, and 4Department of Pharmacy, Johns Hopkins Hospital, Baltimore, Maryland, 21287 USA
`5Division of Pulmonary Medicine, Children’s Hospital of Philadelphia, and Department of Pediatrics,
`University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104, USA
`
`*To whom correspondence and reprint requests should be addressed. Fax: 410-955-1030. E-mail: pzeitli@jhmi.edu.
`
`Most individuals with cystic fibrosis (CF) carry one or two mutations that result in a matura-
`tion defect of the full-length protein. One such mutation, ⌬F508, results in a mutant membrane
`glycoprotein that fails to progress to the apical membrane, where the wild-type protein
`normally functions as a cyclic AMP-regulated chloride channel. 4-Phenylbutyrate (Buphenyl),
`an orally bioavailable short chain fatty acid, modulates heat shock protein expression and
`restores maturation of the ⌬F508 protein in vitro and in vivo. We performed a randomized,
`double-blind, placebo-controlled, dose-escalation and safety study of Buphenyl in 19 adults with
`CF (homozygous ⌬F508) to test the hypothesis that Buphenyl would be safe, well-tolerated,
`and associated with an increase in chloride transport in nasal epithelia. Three dose levels (20,
`30, or 40 g divided t.i.d.) of drug or placebo were given for 1 week. Serial measurements of
`chloride transport by nasal potential difference (NPD) testing and metabolic safety testing were
`performed. A maximum tolerated dose of 20 g was defined based on minimal adverse
`reactions, the safety profile, and a statistically significant induction of chloride transport that
`was maximal by day 3. This short-term phase I/II study demonstrates proof of principle that
`modulation of ⌬F508 CFTR biosynthesis and trafficking is a viable therapeutic approach for
`cystic fibrosis.
`
`Key Words: butyrates, clinical trial, cystic fibrosis, mutation, chloride,
`sodium, sweating, CFTR, ⌬F508, Buphenyl
`
`INTRODUCTION
`The ⌬F508 CFTR mutation results in a mutant membrane
`glycoprotein that fails to progress to the apical membrane,
`where the wild-type protein normally functions as a cyclic
`AMP-regulated chloride channel [1,2]. Instead, the major-
`ity of nascent ⌬F508 CFTR molecules becomes ubiquiti-
`nated and rapidly degraded from the endoplasmic reticu-
`lum [3]. More efficient folding and maturation of ⌬F508
`CFTR can be induced by high concentrations of glycerol
`[4] or by protein assembly at 27⬚C [5,6]. Both HSP70 and
`HSC70, distinct members of the 70 kDa heat shock pro-
`tein family, interact with CFTR, and regulation of these
`heat shock protein–CFTR interactions can restore ⌬F508
`CFTR trafficking [7,8]. We recently demonstrated that 4-
`phenylbutyrate (Buphenyl), an orally bioavailable short
`chain fatty acid, modulates heat shock protein function
`and restores ⌬F508 maturation in vitro and in vivo [7–9]. It
`is not known whether restoration of ⌬F508 CFTR to the
`
`plasma membrane will be sufficient to reverse cystic fibro-
`sis (CF) disease.
`We performed a randomized, double-blind, placebo-
`controlled, dose-escalation, and safety study of Buphenyl
`in 19 adults with CF (homozygous ⌬F508). We hypothe-
`sized that Buphenyl would be safe, well-tolerated, and
`associated with a gain in chloride transport in nasal epithe-
`lia as quantified by nasal potential difference (NPD) test-
`ing. Drug or placebo was administered in three oral dose
`levels (20, 30, or 40 g) divided thrice daily (t.i.d.) for 1
`week. Serial measurements of NPD, sweat electrolyte con-
`centration, metabolic and hepatic function, pulmonary
`function, and sputum microbiology were performed dur-
`ing the study period and during a 1 month washout
`period. Pharmacokinetics were evaluated during the first
`72 hours of study drug administration and will be reported
`separately.
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`1525-0016/02 $35.00
`
`119
`
`LUPIN EX. 1012
`
`

`

`TRIAL
`
`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`RESULTS
`
`Demographics
`There were 12 men and 7 women randomized in the study,
`and all 19 completed the final study visit. Mean age ± SD
`was 28.5 years ± 7.1, average weight was 62.6 kg ± 7.1, and
`average FEV1 (% predicted) was 63.7 ± 17.0. There were no
`significant differences in gender, baseline age, weight, or
`FEV1 among participants in the four groups (Kruskal–Wallis
`test, P > 0.25 for each comparison).
`
`FIG. 1. Baseline NPD. Baseline NPD was recorded on day 0 during superfu-
`sion of Ringer’s solution (triangle = median).
`
`Nasal Epithelial Chloride Transport
`NPD measurements were performed separately in the right
`and left nares. Several subjects presented with inflammation
`and tenderness on one side and NPD was not performed on
`that side if it was painful. Bilateral testing was resumed
`when the symptoms subsided. The baseline NPD did not dif-
`fer between groups (Fig. 1). The means and standard devi-
`ations of the baseline sodium and chloride responses in
`each study group (Table 1) were comparable to data
`obtained in CF subjects in the Cystic Fibrosis Therapeutics
`Development Network (M.P.B. et al., manuscript submit-
`ted), and there were no statistically significant differences
`in these baseline parameters between groups. In CF, the
`basal NPD is determined by the sodium potential which is
`unchecked (< –30 mV) in the absence of functional CFTR.
`Of the intended 38 measured basal NPDs, one naris in each
`of three subjects was too painful to allow measurement at
`the baseline. The baseline NPD was more positive than
`expected (–13 to –25) in 5 of 35 nares. This was due to tech-
`nical difficulties related to patient reports of tenderness at
`the desired point of catheter placement. Although in those
`cases the catheter was repositioned for patient comfort and
`may not have represented the ideal site, these values were
`retained in the analysis. Average baseline NPD was similar
`across groups and indicative of CF.
`The aggregate data for the measured low chloride/iso-
`proterenol response over time of the placebo (Fig. 2A), 20
`g (Fig. 2B), and 30 g (Fig. 2C) cohorts were assessed sepa-
`rately by group. No change was observed for the placebo
`group during the first 7 days. In contrast the low chlo-
`ride/isoproterenol responses increased at days 2, 3, 4, and
`7, respectively, for the 20 g (Fig. 2B) and 30 g (Fig. 2C)
`cohorts. The primary outcome variable, the change in the
`low chloride/isoproterenol response from baseline levels
`on each study day, was then compared across groups for
`days 2 (Fig. 3A), 3 (Fig. 3B), 4 (Fig. 3C), and 7 (Fig. 3D). A
`statistically significant induction of chloride transport
`potentials was observed in a dose-dependent relationship
`on day 3 (Fig. 3B), but this difference diminished by day
`4 (Fig. 3C). The mean difference in low chloride/isopro-
`terenol change between the 20 g cohort and the control
`group was –2.2 mV (95% confidence interval (CI): –10.1,
`
`Adverse Events
`As in our previous study, minor adverse events in the 20
`g cohort included transient nausea, headache, and sleepi-
`ness after the initial dose, and body odor. The first three
`resolved with a dose of Tylenol, and hydration was encour-
`aged. Body odor was an inconsistent complaint by family
`or friends of subjects. No dose adjustments were required.
`These complaints were also observed after the initial dose
`in the 30 g cohort. Several subjects reported visual distur-
`bances that were transient after the first dose. One subject
`had severe headache that resolved with a reduction to 20
`g daily. All three subjects in the 40 g cohort complained
`of nausea, headache, and visual disturbances, and one
`complained of cramps in the hands and fingers. One of
`these subjects tolerated 40 g of Buphenyl when it was
`divided into six daily doses. One tolerated a reduction to
`30 g daily, and one subject found the symptoms to be so
`unpleasant that the drug had to be discontinued. The data
`and safety monitoring committee
`was convened to review the
`adverse event profile and recom-
`mended termination of the 40 g
`cohort. Although the maximum
`tolerated dose was 30 g daily, the
`number of tablets necessary and
`the side effect profile, as well as
`the physiologic outcome, suggest
`that the practical daily dose is 20
`g daily divided t.i.d.
`
`Group
`
`n
`
`Amiloride
`inhibition (mV)
`22.3 (9.7)
`7
`Control
`23.9 (9.1)
`11
`20 g
`25.4 (14.6)
`11
`30 g
`27.2 (7.8)
`6
`40 g
`Data are expressed as mean (SD).
`
`TABLE 1: Basal NPD parameters
`Low chloride
`Isoproterenol
`response (mV)
`response (mV)
`3.9 (4.3)
`0.1 (1.6)
`4.4 (7.2)
`3 (2.0)
`4.5 (6.8)
`1.4 (2.9)
`–2 (4.1)
`3.7 (4.6)
`
`Low chloride/isoproterenol
`response (mV)
`4 (4.2)
`7.5 (8.0)
`5.9 (5.7)
`1.7 (5.5)
`
`120
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`
`LUPIN EX. 1012
`
`

`

`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`TRIAL
`
`A
`
`C
`
`5.8) and –11.9 mV (95% CI: –18.9, –5.0) on days 2 and 3,
`respectively. This difference decreased to –1.8 mV (95% CI:
`–10.0, 6.3) by day 4. A similar trend was observed in the
`comparison of the low chloride/isoproterenol change
`between the 30 g cohort and the control group from base-
`line to day 2 (–5.0 mV (95% CI: –14.3, –4.2)) and to day
`3 (–6.3 mV (95% CI: –16.7, 4.1)).
`Analysis of the isoproterenol response alone, without
`the previous low chloride maneuver, revealed the same
`patterns. A statistically significant, dose-dependent
`improvement of isoproterenol-stimulated chloride trans-
`port was observed on days 2 and 3, but this difference
`diminished by day 4. The mean difference in isoproterenol
`change between the 20 g cohort and the control group was
`–5.6 mV (95% CI: –10.2, –0.9) and –6.2 mV (95% CI: –12.1,
`–0.3) on days 2 and 3, respectively. This difference
`decreased to –2.8 mV (95% CI: –10.3, 4.7) by day 4. A sim-
`ilar trend was observed in the comparison of the difference
`in isoproterenol change on day 2 between the 30 g cohort
`and the control group (–5.6 mV (95% CI: –9.4, –1.7))
`although the difference was not statistically significant by
`day 3 (–2.6 mV (95% CI: –8.9, 3.7)).
`
`B
`
`FIG. 2 Nasal epithelial chloride transport during treatment and washout peri-
`ods. Individual data points representing the low chloride/isoproterenol
`responses at each study date during treatment are recorded by +. (A) Control
`group; (B) 20 g cohort; (C) 30 g cohort. Note that the induction of chloride
`transport in the phenylbutyrate treated cohorts is indicated by negative val-
`ues crossing below the zero line.
`
`There was no statistically significant change from base-
`line in the amiloride-inhibited potential at any time point
`(data not shown). Thus, induction of chloride transport
`was not accompanied by inhibition of amiloride-regulated
`sodium transport as would be predicted for full correction
`of CFTR function. One example of a robust induction of
`chloride transport in an individual in the 30 g cohort on
`day 3 is shown (Fig. 4A). This NPD measurement tracing
`demonstrates the persistently high basal level of NPD in
`spite of normal levels of chloride transport, which again
`suggests that induction of chloride transport has no sig-
`nificant effect on amiloride-regulated sodium transport.
`Urine and plasma collected on day 7 demonstrated the
`presence of study drug metabolite in all treated subjects
`and not in controls. Phenylbutyrate was efficiently con-
`verted to phenylacetate and excreted in the urine as
`phenacetylglutamine in the 20 g cohort. Increasing the
`dose by 33% from 20 g to 30 g daily was associated with
`a doubling in the AUC24 (11.9 ± 5.9 and 22.6 ± 5.0 mm*h).
`Examples of the plasma phenylbutyrate (Fig. 4B) and
`phenylacetate (Fig. 4C) profiles for the same subject dis-
`played in Fig. 4A are given. These graphs demonstrate the
`three expected daily peaks in phenylbutyrate concentra-
`tion and more sustained levels of the first metabolite dur-
`ing the waking hours.
`Accumulation of phenylacetate in the plasma was
`observed in one individual in the 30 g cohort. This sug-
`gests that in this subject, phenylbutyrate may have satu-
`rated the metabolic pathway to conversion to phenacetyl-
`glutamine, thus suggesting a maximum tolerated dose of
`20 g daily.
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`
`121
`
`LUPIN EX. 1012
`
`

`

`TRIAL
`
`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`A
`
`C
`
`B
`
`D
`
`FIG. 3. Change from baseline in nasal epithelial chloride transport (triangle = median). Nasal epithelial chloride transport was measured on day 2 (A), day 3 (B),
`day 4 (C), and day 7 (D). Days 2 and 3 resulted in maximal inductions of chloride transport for 20 and 30 g cohorts.
`
`Sweat Electrolytes
`Although several individuals exhibited a decrease in sweat
`chloride while taking the study drug with restoration of
`pre-drug values during the washout, there was significant
`inter-subject variability and no statistically significant dif-
`ference for the groups as a whole (Table 2).
`
`Hepatic Enzymes
`One subject had an isolated elevation in alkaline phos-
`phatase at baseline which persisted. There was a trend
`toward reduction in alanine aminotransferase (ALT) and
`aspartate aminotransferase (AST) in the 20 g and 30 g
`cohorts and a statistically significant drop in total biliru-
`bin by day 7 for the 30 g cohort (Table 2).
`
`Uric Acid
`Uric acids levels became mildly elevated while on study
`drug in the 30 g cohort and returned to baseline levels
`
`during the washout period (Table 2). Phenacetyl-
`glutamine and urate may compete for the same trans-
`porter in the kidney.
`
`Pulmonary Function Test
`There was no statistically significant difference from base-
`line in FEV1 levels beetween groups, according to meas-
`urements made on day 0, 3, 4, or 7 within each group.
`
`Microbiology
`Pseudomonas aeruginosa and Staphylococcus aureus were
`scored semi-quantitatively at baseline and day 7 using a
`Likert scale from 0 (none) to 6 (heavy). Median score for
`P. aeruginosa at baseline was 4.5 in the controls, 2 in the
`20 g cohort, and 5 in the 30 g cohort. There were no sta-
`tistically significant differences in these scores at the base-
`line measurement between groups or between baseline
`and day 7 within each group.
`
`122
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`
`LUPIN EX. 1012
`
`

`

`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`TRIAL
`
`A
`
`B
`
`C
`
`DISCUSSION
`
`Magnitude of Chloride Response in
`CF Nasal Epithelia
`The combined low chloride/isoproterenol response in the
`NPD test has been recognized to be the most discrimina-
`tory measure of CFTR dysfunction and cystic fibrosis [10].
`Homozygous ⌬F508 subjects (uniformly pancreatic insuf-
`ficient) would be expected to have negligible, or at most,
`extremely low chloride responses (< –5 mV low
`
`FIG. 4. Example of a NPD in a subject taking 30 g phenylbutyrate
`daily for 3 days. (A) NPD. (B) Plasma phenylbutyrate levels for
`study days 1–3. (C) Plasma phenylacetate levels for study days
`1–3.
`
`chloride/isoproterenol responses). CF can be
`associated with slightly higher or intermediate
`levels of low chloride/isoproterenol response,
`particularly in the setting of one or more pan-
`creatic sufficient mutations [11–14]. In our
`study, restoration of the ⌬F508 CFTR to the
`plasma membrane may have been associated
`with a sub-maximal response because this muta-
`tion carries a shorter open time and lower con-
`ductance than wild-type CFTR [15]. We had previously
`observed, in subjects of this genotype taking 19 g daily
`who participated in a pilot study, a modest improvement
`of low chloride/isoproterenol response [8]. In the present
`study, a dose escalation was undertaken to attempt to
`increase the low chloride/isoproterenol response above
`what we observed in our first pilot clinical trial using 19
`g daily. Whereas some subjects achieved low chloride/iso-
`proterenol responses approaching normal values, others
`did not. Our results suggest additional genes or individ-
`ual variation in pharmacokinetics may play a role in the
`response, but that on average, we have achieved a maxi-
`mal response using 20 g daily. Our data also indicate that
`there was no additional advantage to taking 30 g daily.
`We observed peak improvement of the low chlo-
`ride/isoproterenol response between 3 and 4 days,
`depending on the dose, and a slight diminution in this
`response by day 7 in all dose groups. While the molecu-
`lar detail of this observation is not known, two potential
`explanations are supported by published observations.
`Both Linsdell [16], in single channel recordings, and
`Loffing et al. [17], in Calu-3 cells, observed inhibition of
`CFTR-mediated Cl– transport with millimolar concentra-
`tions of 4-phenylbutyrate (4PBA), suggesting direct inhi-
`bition of the channel by 4PBA. Inhibition of CFTR by
`4PBA in single channel recordings was only observed with
`application of 4PBA to the cytoplasmic face of the chan-
`nel, and not with application of 4PBA to the extracellu-
`lar face. For this mechanism to apply in vivo, a significant
`and sustained intracellular accumulation of Buphenyl at
`millimolar concentrations would need to occur. Only
`transient plasma concentrations of Buphenyl greater than
`1 mM (Fig. 4B) are observed. Rapid metabolism of
`Buphenyl is also observed in patients with urea cycle dis-
`orders, and our pharmacokinetic data (reported else-
`where), reporting quantitative conversion of Buphenyl to
`phenylacetate and phenacetylglutamine before excretion
`in the urine, argues against such a mechanism.
`Furthermore, patients with urea cycle disorders who have
`taken Buphenyl daily for many years have not developed
`lung disease, bronchiectasis, or phenotypic features of CF
`(Saul Brusilow, April 2002, pers. comm.).
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`
`123
`
`LUPIN EX. 1012
`
`

`

`TRIAL
`
`Test
`Sweat chloride
`– baseline
`- Day 2
`- Day 3
`- Day 4
`- Day 7
`
`Uric acid –
`baseline
`- Day 2
`- Day 3
`- Day 4
`- Day 7
`
`AST – baseline
`- Day 2
`- Day 3
`- Day 4
`- Day 7
`
`ALT – baseline
`- Day 2
`- Day 3
`- Day 4
`- Day 7
`
`TABLE 2: Secondary outcome measures
`Control
`20 g
`Mean (SD)
`Mean (SD)
`110.1 (16.6)
`119.2 (12.7)
`
`119.2 (7.4)
`113.6 (13.6)
`111.0 (8.7)
`117.2 (10.3)
`
`119.8 (10.4)
`117.2 (9.0)
`116.2 (9.9)
`116.5 (10.7)
`
`6.3 (1.0)
`
`6.3 (2.0)
`6.0 (1.4)
`5.9 (1.4)
`5.7 (1.6)
`
`45.0 (41.0)
`27.5 (13.2)
`29.2 (13.8)
`47.0 (34.0)
`59.0 (55.2)
`
`49.5 (42.5)
`40.0 (33.4)
`41.2 (30.0)
`56.2 (52.8)
`81.5 (78.5)
`
`5.6 (1.8)
`
`6.7 (1.9)
`6.6 (1.6)
`6.8 (1.6)
`7.4 (2.0)
`
`23.5 (3.0)
`20.2 (4.5)
`19.2 (3.1)*
`19.5 (3.7)*
`22.4 (6.3)
`
`28.5 (14.4)
`23.7 (7.1)
`21.0 (7.7)*
`19.2 (4.4)*
`19.4 (5.9)*
`
`30 g
`Mean (SD)
`116.8 (16.8)
`
`96.2 (45.3)
`121.1 (13.5)
`110.3 (16.2)
`119.8 (4.4)
`
`6.8 (0.9)
`
`8.7 (1.7)*
`9.4 (2.5)*
`8.9 (2.6)
`8.6 (2.1)
`
`28.3 (8.1)
`27.2 (8.2)
`23.0 (8.5)
`38.0 (39.9)
`33.7 (18.6)
`
`35.0 (14.5)
`32.6 (21.5)
`26.2 (18.5)*
`25.7 (15.1)*
`42.5 (74.8)
`
`238.2 (209.8)
`
`98.2 (32.9)
`
`162.8 (67.8)
`
`Alkaline phosphatase –
`baseline
`- Day 2
`- Day 3
`- Day 4
`- Day 7
`
`Total bilirubin –
`baseline
`- Day 2
`- Day 3
`- Day 4
`- Day 7
`
`228.0 (203.8)
`221.2 (194.4)
`216.0 (190.7)
`353.5 (277.9)
`
`102.8 (41.4)
`95.5 (35.2)
`92.0 (32.4)
`96.0 (15.0)
`
`0.5 (0.3)
`
`0.5 (0.2)
`0.4 (0.2)
`0.5 (0.1)
`0.4 (0.1)
`
`0.6 (0.4)
`
`0.8 (0.7)*
`0.7 (0.4)*
`0.6 (0.3)
`0.5 (0.2)
`
`*Significant change from baseline, P < 0.05.
`
`We favor a second possible explanation for the
`diminution of the low chloride/isoproterenol response
`between 4 and 7 days, namely that the cell adapts to the
`perturbation of cellular homeostasis caused by Buphenyl
`that initially improved ⌬F508 trafficking. Such adapta-
`
`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`tion returns the cell closer to its base-
`line condition in which ⌬F508 traffick-
`ing is aberrant. Such a mechanism is
`consistent with
`suggestions
`that
`Buphenyl alters molecular chaperone
`expression and interaction with ⌬F508
`[7,18] and causes a more global cellular
`adaptation, one element of which is
`increased turnover of HSC70 mRNA
`[19]. Such global effects on the cell may
`result from Buphenyl’s action as an
`inhibitor of histone deacetylase [20]
`and are also consistent with a long-
`term loss of Buphenyl’s effect on
`increased hemoglobin F (HbF) expres-
`sion in patients with ␤-hemoglo-
`binopathies.
`Upregulation of these intermediate
`levels of chloride transport was not asso-
`ciated with a reduction in the basal NPD
`nor with a reduction in the amiloride
`response, although the presence of func-
`tional ⌬F508 in the plasma membrane
`might have been expected to downreg-
`ulate the epithelial sodium channel [21].
`Because
`some members of
`the
`isoflavonoid family have been shown to
`activate chloride conductance through
`CFTR, it may be possible to increase
`⌬F508-mediated chloride transport by
`adding molecules that increase CFTR
`open time, decrease CFTR closed time,
`or prolong CFTR residence in the plasma
`membrane by inhibiting recycling.
`Perhaps full activation of ⌬F508-medi-
`ated chloride transport would downreg-
`ulate basal NPD and amiloride-inhibited
`sodium transport, as is suggested by
`Suaud et al. [22].
`
`149.6 (65.3)*
`147.0 (63.3)*
`120.8 (58.1)*
`144.5 (74.8)
`
`0.5 (0.2)
`
`0.5 (0.2)
`0.4 (0.2)
`0.7 (0.7)
`0.3 (0.1)*
`
`Systemic Administration
`The ability to administer a butyrate
`through the gastrointestinal tract is a
`distinct advantage, because CF is a sys-
`temic disorder of exocrine glands and
`secretory epithelia. Sinusitis, pancreati-
`tis, biliary tract disease, hepatitis, meco-
`nium ileus equivalent, and sweat losses
`leading to dehydration should theoreti-
`cally respond to restoration of ⌬F508
`CFTR trafficking. Our study was not suf-
`ficiently powerful statistically to detect a difference in the
`secondary outcomes of sweat electrolytes, hepatic
`enzymes, or FEV1 of subjects. Importantly, we did not
`observe an apparent worsening of any of these measures
`during phenylbutyrate treatment.
`
`124
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`
`LUPIN EX. 1012
`
`

`

`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`TRIAL
`
`Tolerability
`Phenylbutyrate is approved for infants, children, and
`adults with urea cycle disorders that lead to accumulation
`of nitrogen in the form of ammonia. Therapy must be
`daily and lifelong. In general, it is well-tolerated and asso-
`ciated with improvements in ammonia and liver function.
`Phenylbutyrate also has been tested in phase I and II clin-
`ical trials for a variety of solid tumors [23]. Some subjects
`have remained on extremely high doses of the drug for
`more than a year. Side effects are similar to the ones
`observed in our trial—mainly headache and nausea.
`Phenylbutyrate also has undergone testing in the hemo-
`globinopathies [24,25]. The therapy is pulsatile and life-
`long. Identification of the appropriate dosage regimen for
`CF as well as potential synergy with channel activators
`should be explored.
`
`Genetics
`Upregulation of cellular ⌬F508 CFTR protein is desirable
`because the disease results from a deficiency of CFTR
`beyond the endoplasmic reticulum. Upregulation of
`other mutations which create a partially functional CFTR
`is also a desirable goal. Presumably a larger fraction of
`modestly impaired CFTR proteins in the plasma
`membrane would still improve salt and water balance
`across the epithelium overall. Phenylbutyrate has not
`been tested with any of the additional > 900 mutations
`seen in CF.
`
`Applicability in Other Inherited Disorders
`The ⌬F508 CFTR is prematurely degraded from the endo-
`plasmic reticulum. Accumulation of undigested, mis-
`folded CFTR in vacuoles retrieved from the endoplasmic
`reticulum is not a key feature of CF. The zz genotype of
`␣
`1anti-trypsin (␣
`1AT) deficiency produces a misfolded
`␣
`1AT that is accumulated in the hepatocyte, the cell of
`origin, causing progressive cirrhosis in a subgroup of
`patients. The chaperones involved in ␣
`1AT biogenesis are
`similar to those involved in CFTR processing. We have
`implicated butyrate-mediated changes in HSC70–CFTR
`and HSP70–CFTR complexes as key events in the upreg-
`ulation of CFTR and chloride transport. Similar mecha-
`nisms may account for zz ␣
`1AT secretion in the murine
`model [26]. Phenylbutyrate may therefore be helpful to
`the
`subset of CF
`subjects presenting with
`significant liver involvement.
`
`METHODS
`Human subjects. Patients of both genders with CF who were homozy-
`gous ⌬F508 genotype and ±18 years of age were eligible to participate.
`Exclusion criteria included pregnancy or inability to practice birth con-
`trol, breast-feeding, clinically significant liver disease, 24-hour oxygen
`therapy, acute upper respiratory infection, acute pulmonary exacerba-
`tion, and sinus surgery within 6 weeks or intravenous administration of
`antibiotics within 4 weeks of the start of the trial. All subjects signed
`informed consent forms. The study protocol and consent forms were
`
`approved by the Joint Committee on Clinical Investigation of the Johns
`Hopkins School of Medicine (IRB).
`Nineteen subjects entered the study and were randomized. The study
`drug was discontinued in one female subject from the 30 g cohort who
`experienced acute distal intestinal obstruction syndrome on day 2, and this
`subject was replaced by another subject. Due to unpleasant symptoms,
`three subjects required a dose reduction; two subjects from the 30 g cohort
`were reduced to 20 g and one from the 40 g cohort was reduced to 30 g
`daily. One subject in the 40 g cohort elected to discontinue use of the
`study drug due to intolerable symptoms. Randomization in the 40 g cohort
`was discontinued at the request of the data and safety monitoring com-
`mittee after the third subject experienced unacceptable symptoms. This
`resulted in six subjects randomized to 20 g, six randomized to 30 g, three
`randomized to 40 g, and four randomized to placebo. All subjects com-
`pleted the final visit.
`Study design. A randomized, double-blind, dose-escalation design was
`used. Subjects were randomized in a 3:1 fashion to either the study drug
`or placebo. The designed study included eight subjects randomized at each
`of three sequential dose levels resulting in six subjects per dose cohort and
`six subjects in the placebo group. Escalation to the subsequent dose level
`proceeded unless unacceptable toxicity was observed. The primary out-
`come was the change in NPD between the baseline measurement and meas-
`urements taken on day X of study drug administration. A sample size of
`six subjects per group was planned with a power of 0.80 and a two-sided
`significance level of 0.05 to detect a difference in mean change in NPD of
`5 mV between drug and placebo, assuming a standard deviation of the
`change of 1.1 mV in placebo and 2.9 mV in drug. These estimates were
`based on our previous investigation of the isoproterenol/low chloride
`response in homozygous ⌬F508 subjects [8]. Secondary outcomes included
`changes in metabolic and hepatic function, pulmonary function, sputum
`microbiology and pharmacokinetics.
`Study drug and placebo. Buphenyl (sodium salt of 4-phenylbutyrate) in
`a 500 mg tablet was provided by Medicis Pharmaceutical Inc. A sodium lac-
`tate placebo tablet was formulated to maintain equivalent sodium salt load
`by Medicis Pharmaceutical Inc. A “no added salt” diet was prescribed for
`the duration of study drug administration. The 20 g daily dose was divided
`into 13 tablets to be taken at 8 AM and at 2 PM and 14 tablets to be taken
`at 8 PM. The 30 g daily dose was divided into 20 tablets to be taken at 8
`AM, at 2PM, and at 8 PM. The 40 g daily dose was first prescribed as 27
`tablets to be taken at 8 AM and at 2 PM and 26 tablets to be taken at 8 PM.
`This latter schedule proved unpleasant, and the doses were then divided
`into 14 tablets to be taken at 8 AM, 13 tablets to be taken at 11 AM, 14
`tablets to be taken at 2 PM, and 13 tablets to be taken at 5 PM, at 8 PM,
`and at 11 PM. The dose was tolerated better when administered with food.
`Hydration, either by the oral route or through the intravenous line used
`for blood sampling, was associated with fewer symptoms of nausea and
`headache, which resolved in the first 24 hours.
`Study protocol. Subjects were admitted to the Johns Hopkins Pediatric
`General Clinical Research Center (GCRC) for the first 4 days of study drug
`administration to monitor safety, to obtain frequent blood sampling and 24
`hour urine collection to evaluate the pharmacokinetics, and to perform daily
`measurements of CFTR function. Routine safety tests consisted of measure-
`ment of hematologic indices, prothrombin time (PT) and partial thrombo-
`plastin time (PTT), hepatic function, urinalysis, spirometry, and electrolyte
`and uric acid levels. Blood sampling through a heparin lock or midline
`catheter was obtained at 0, 1, 2, 4, 6, and 8 hours post-dose on days 1, 2,
`and 3. Twenty-four hour urine collections were conducted on days 1, 2, and
`3. Sputum was collected at baseline, on day 7, and weekly during the remain-
`der of the protocol (1 month washout period). Sweat chlorides were meas-
`ured in the Johns Hopkins Hospital chemistry lab on sweat collected by pilo-
`carpine electrophoresis from equivalent sites on each forearm at baseline, and
`on days 2, 3, 4, and 7, and at weekly intervals during the washout period.
`NPD testing as described [8] (with minor modifications) was performed in
`both nares at baseline, on days 2, 3, 4, and 7, and weekly during the washout
`period. Testing was performed with room temperature solutions, Ringer’s
`solution instead of an agar bridge for the subcutaneous electrode, and com-
`puter-based data acquisition for the last 5 patients.
`
`MOLECULAR THERAPY Vol. 6, No. 1, July 2002
`Copyright © The American Society of Gene Therapy
`
`125
`
`LUPIN EX. 1012
`
`

`

`TRIAL
`
`doi:10.1006/mthe.2002.0639, available online at http://www.idealibrary.com on IDEAL
`
`Analysis of metabolites by HPLC. Plasma (0.1 ml) was treated with 0.3
`ml methanol containing 0.1 mM benzoyl glycine methyl ester as an inter-
`nal standard. After 3 minutes at room temperature, this sample was cen-
`trifuged in a microfuge at top speed for 5 minutes. The supernatant was
`analyzed for metabolites. Urine (0.25 ml) was prepared with 0.75 ml
`methanol containing the standard. A 1:25 dilution (volume:volume) of
`the supernatant was analyzed on a Waters Alliance system using a Waters
`C18 Bondapak column, 3.9 mm ⫻ 300 mm. The column temperature was
`maintained at 40⬚C, and the flow rate was 1.5 ml/min. The mobile phase
`consisted of (A) 1 mM phosphate, pH 3; (B) methanol; and (C) 1 mM phos-
`phate, pH 4. Phenylbutyrate, phenylacetate, and phenacetylglutamine were
`resolved and quantified by A220 measurement.
`Statistical methods. Data were graphically displayed with exploratory
`techniques including dot plots and box and whisker plots. Differences in
`baseline characteristics between groups were assessed using the
`Kruskal–Wallis nonparametric analysis of variance method [27]. Changes
`in NPD, between baseline and treatment day within groups, were initially
`assessed using the Wilcoxon signed-rank test [27]. Changes in NPD over
`time were analyzed by a regression approach using a robust estimate of vari-
`ance to adjust for the correlation between measures on different nares of
`the same individual [28].
`
`ACKNOWLEDGEMENTS
`P.L.Z. was supported by Cystic Fibrosis Foundation grants ZEITLI97A0, P01 HL
`51811, and NCRR RR00052 for this work. M.D.W. received support from the Cystic
`Fibrosis Foundation. R.C.R. was supported by Cystic Fibrosis Foundation grant
`RUBENS96LO. M.P.B. was supported by a Cystic Fibrosis Foundation Shw/achman
`Award. A licensing agreement exists among the Johns Hopkins University, Ucyclyd
`Pharma, Inc. (Medicis), and P.L.Z. The terms of this arrangement are being man-
`aged by the University in accordance with its conflict of interest policies.
`
`RECEIVED FOR PUBLICATION FEBRUARY 20; ACCEPTED MAY 8, 2002.
`
`3.
`
`REFERENCES
`1. Kartner, N., Augustinas, O., Jensen, T. J., Naismith, A. L., and Riordan, J. R. (1992).
`Mislocalization of ⌬F508 CFTR in cystic fibrosis sweat gland. Nat. Genet. 1: 321–327.
`2. Cheng, S. H., et al. (1990). Defective intracellular transport and processing of CFTR is
`the molecular basis of most cystic fibrosis. Cell 63: 827–834.
`Jensen, T. J., et al. (1995). Multiple proteolytic systems, includin