INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: emphysema; all-trans-retinoic acid; clinical trial; pulmonary function tests; X-ray CAT scan

The pathogenesis of emphysema is complex, most often resulting from tobacco-induced inflammation and/or chronic elastase-induced destruction of alveolar tissue (1). This tissue destruction reduces the surface area of the lung available for gas exchange and reduces elastic recoil, ultimately leading to hyperinflation, airflow limitation, ventilation/perfusion mismatching, impaired oxygen transfer, increased pulmonary vascular resistance, and right heart dysfunction. Unfortunately, the adult lung cannot spontaneously regenerate itself, and the destructive effects of emphysema have heretofore been considered progressive and irreversible. In the absence of a cure, medical therapies have focused on preserving the remaining lung, reducing complications, or controlling and ameliorating symptoms (2). As of now, no available treatment has directly addressed the goal of regenerating functional lung tissue.

Retinoids, such as vitamin A and all-trans-retinoic acid (ATRA), are known to activate genes involved in lung development and promote alveolar septation and growth in the pre- and postnatal period (3, 4). In a recent animal model, systemic administration of ATRA reversed manifestations of elastase-induced emphysema (5). In that study, elastase was instilled into the lungs of mature rats to produce changes characteristic of emphysema. After allowing 25 d for the destructive changes to stabilize, a 12-d treatment with ATRA resulted in regeneration of damaged lung with significant reversal of both the anatomic and physiologic manifestations. No spontaneous regeneration was observed in the absence of retinoid treatment. Similar results have now been replicated by Belloni and coworkers (6) with both ATRA and 9-cis-retinoic acid.

Although these findings in rats are exciting, caution is warranted in drawing too close a parallel between this elastase-induced model of emphysema and the pathogenesis of human emphysema. Unlike this model, where unaffected lung tissue remained relatively normal, the remaining alveolar tissue in patients with emphysema is functionally and structurally altered. In addition, interspecies and age-related differences in gene regulation exist. Therefore, well-controlled human clinical trials are necessary to understand the effects of retinoids on the mature human lung.

We report on a pilot study evaluating ATRA for its application as a medical treatment for human emphysema. ATRA is a Food and Drug Administration (FDA)-approved orphan drug recognized for its activity against acute promyelocytic leukemia (APL) (7, 8). ATRA has also been tested experimentally for its utility as a chemopreventive agent and an antineoplastic for the treatment of solid tumors, and for a variety of dermatologic uses (9, 10). Our goal was to determine whether this drug could be safely administered to subjects with advanced emphysema and whether standard measures of pulmonary function, noninvasive assessment of lung pathology, and respiratory-related symptoms could be used to identify potential benefits. We were also interested in the pharmacokinetics of drug administration when prescribed as a chronic therapy. Our results suggest that ATRA is well tolerated and that further studies are feasible.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design and Medication

A double-blind, placebo-controlled crossover design was employed, in which patients were randomized at a 1:1 ratio to receive either 12 wk of ATRA (Vesanoid^®; Roche Laboratories, Inc., Nutley, NJ) followed by 12 wk of a matching placebo, or 12 wk of placebo followed by 12 wk of ATRA. Gelcaps (10 mg) were consumed twice a day at a dose of 25 mg/m²/dose (50 mg/m²/d) for four consecutive days out of every week. This intermittent dosing regimen was employed as a strategy to prevent the down-regulation of ATRA blood levels that occurs with continuous administration (11). All subjects continued on standard medical therapy for emphysema at the direction of their primary physician.

Subjects and Screening

Written informed consent was obtained according to guidelines established by the UCLA Institutional Review Board. Patients aged 45 to 80 yr, with a clinical history of moderate to severe emphysema, were eligible if they met two of three pulmonary function criteria (FEV₁ < 60% predicted, total lung capacity (TLC) >110% predicted, or diffusing capacity of the lung for CO (DL_CO) < 60% predicted) and demonstrated visual evidence of emphysema on computed tomography (CT) scan. All subjects were required to be nonsmokers for at least 6 mo and demonstrate normal renal, bone marrow, and hepatic function. Exclusion criteria included evidence of resting hypoxemia (O₂ saturation < 90%); significant airway hyperreactivity (FEV₁ increase > 30% of baseline or 300 ml after bronchodilators); FEV₁ < 800 ml and 30% predicted; Karnofsky performance score < 65; unexplained weight loss or lung nodules; concurrent pulmonary conditions (e.g., giant bulla[e], bronchiectasis, pneumonia, interstitial lung disease, cancer, or pulmonary embolism); hyperlipidemia (cholesterol > 220 mg% and/or triglyceride > 300 mg%); active cardiac diseases or hepatitis; or corticosteroid dependency (more than 2 wk out of every 2 mo). Concurrent medications known to alter cytochrome P-450 function were excluded during the study period. Women of childbearing potential were excluded due to the teratogenic effects of ATRA.

Monitoring and Outcome Measures

Outcome measures, including pulmonary function testing (PFTs), chest CT, and quality of life (QOL) questionnaires were obtained at baseline and at 3-mo intervals. Patients were evaluated at 2 wk after the start of each phase and then monthly with history and physical examinations, complete blood counts, and a chemistry panel. Plasma samples for determination of ATRA levels were obtained before and 3 h after the initial dose of medication, and then 3 h after the first dose of Weeks 3 and 8 of each phase.

Pulmonary function testsGlobal physiologic responses were measured by serial PFTs including spirometry, lung volumes by body plethysmography, static transpulmonary pressure-volume curves, and DL_CO, using automated pulmonary function equipment (6200 Autobox DL; SensorMedics, Yorba Linda, CA). American Thoracic Society (ATS) criteria for standardization of spirometry (12) and DL_CO (13) were used. Subdivisions of lung volume were determined by constant-volume whole-body plethysmography using the panting technique of DuBois and coworkers (14). Arterial blood gases were obtained on room air prior to PFT determinations.

Lung compliance pressure-volume curvesTranspulmonary pressures were determined with the subject seated using a latex esophageal balloon positioned in the lower third of the esophagus (15) and connected to a differential pressure-transducer (Validyne, Northridge, CA; Model MP-45-3). Transpulmonary pressures, volume, and flow were recorded and analyzed via Vmax software (Autobox 6200, SensorMedics).

Chest CT scansNoncontrast CT studies were performed on a dedicated Electron Beam CT Scanner (Imatron Inc., San Francisco, CA) using a standardized imaging protocol. A single breathhold study was obtained at suspended TLC using an 8-mm collimation (thick section), 8-mm table-feed, continuous volume acquisition protocol. High-resolution CT (HRCT) images (1.5 mm) were obtained every 10 mm at TLC and every 20 mm at suspended TLC RV, respectively. To ensure reproducibility of the breathhold, patients were carefully trained using spirometric guidance. All images were transferred electronically to the Thoracic Radiology Imaging and Data Analysis (TRIAD) platform. Quantitative image analysis (QIA) was performed using a previously validated automated lung segmentation and quantitation program (16) to assess lung density and total lung volume. Density masks were generated using 910 Hounsfield units (HU) as a previously validated threshold indicative of emphysema (17). The same thoracic radiologist reviewed all scans for visual confirmation of emphysema and to exclude ancillary pathology.

Quality of Life Questionnaires.QOL was assessed by the St. George's Respiratory Questionnaire (SGRQ), which assessed subjective measures of 76 items divided into three sections: symptoms, activity, and impact (18). A score for each section and a score for the summed weights of these scores were determined and expressed as percentage or points, where 100 denotes very poor health and 0 very good health.

Side effects and toxicityAdverse event questionnaires were administered at each study visit and responses were graded according to a modification of the National Cancer Institute (NCI) Common Toxicity Criteria Scale.

Measurement of plasma ATRA levelsHeparinized plasma samples were collected in foil-covered tubes, processed under yellow light, and stored in amberized vials at 80° C. Analyses were carried out by Cedra Corporation (Austin, TX) using a good laboratory practice-validated liquid chromatography/mass spectroscopy assay.

Statistical Analysis

Changes in pulmonary function at 3 and 6 mo were evaluated using a Wilcoxon's signed rank test. To identify treatment-related effects, changes in the cohort receiving ATRA for the first 3 mo were compared with changes in subjects receiving placebo for the first 3 mo using a two-sample rank sum test. Global CT lung volume and density mask assessments were performed for each lung separately (R versus L), as well as density mask assessment for two representative thick section slices within each lung in the upper and lower lobes. The percentage change in density within lungs or regions of interest were compared before and after each treatment period. Responses to the QOL questionnaires were examined with a Fisher exact test to compare the percentage of subjects with clinically significant changes in their QOL scores in one treatment group with the other treatment group. Statistical analyses were performed using STATA Statistical Software Release 6.0 (Stata Corporation, College Station, TX) and two-tailed probabilities were used throughout.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Enrollment and Baseline Characteristics

During the 1.5-yr study period, approximately 2,500 phone inquiries were received, and 150 subjects forwarded records for review. Of these, 30 were invited for full screening and 20 met all study criteria and were enrolled. The average subject age was 66 yr (range 47-78) and all subjects were former tobacco smokers who had quit smoking at least 6 mo prior to enrollment, including two subjects with ₁-antitrypsin deficiency. Only one patient had symptoms consistent with chronic bronchitis. Patient demographics and baseline studies are summarized in Table 1. Overall, FEV₁ averaged 1.24 L (43% predicted), TLC averaged 7.95 L (129% predicted), and mean DL_COcorr was 1.12 ml/min/mm Hg (36% predicted). According to the SGRQ, respiratory symptoms and their impact were uniformly reported as moderate to severe. No significant differences in PFTs or QOL scores were observed between subjects randomized into Arm 1 versus Arm 2. CT confirmed heterogeneously distributed emphysema of a moderate to severe grade in all cases.

Side Effects and Adverse Events

All randomized subjects completed both phases without evidence of dose-limiting toxicity. In general, ATRA was well tolerated, with the most common side effects including skin changes (dryness, cracking lips, chelitis), mild headache, hyperlipidemia, mild pruritis, mild transaminase elevations, mild muscle and bone pain/soreness, and a sensation of clogged ears (Table 2). These side effects developed primarily during active treatment. All side effects were rated as grade 1 according to NCI criteria except for one elevation in triglycerides greater than two times baseline (grade 2 toxicity) that spontaneously resolved. Twelve of 20 patients had self-limited symptoms of upper respiratory illness (URI) (< 3 wk) during the 6 mo treatment period. In 9 of these 12 patients, URI symptoms occurred during active drug phase, and in 4 of the 12, symptoms occurred during both active drug and placebo phases.

Plasma ATRA Levels

On average, peak plasma drug levels were highest after the first dose of medication and decreased by 25-30% on follow-up testing at Weeks 3 and 8 of the 12-wk treatment (Figure 1A). However, marked variability was observed with respect to both the initial 3-h peak values and the stability of ATRA levels over time. Three general patterns were observed (Figures 1B-1D). Six subjects (30%) exhibited increasing plasma ATRA levels over time. In five patients (25%), levels increased at 3 wk, but either stabilized or decreased by 8 wk. In the remaining nine participants (45%), ATRA levels were highest after the first dose of medication and were significantly lower at all of the follow-up determinations. Overall, despite the intermittent administration of drug for only 4 d each week, seven patients (35%) developed greater than 50% reduction of their 3-h plasma ATRA level at either Weeks 3 or 8 of treatment, whereas five patients had between a 25% and 49% reduction in their follow-up ATRA levels. Plasma ATRA levels for all subjects during the placebo phase ranged from < 2 to 3.75 ng/ml (data not shown).

View larger version (22K): [in this window] [in a new window]   Figure 1.   Average plasma ATRA levels decline over the course of treatment but with considerable variation between subjects. Plasma ATRA levels were determined by liquid chromatography/mass spectroscopy assay at baseline and again at 3 h after the first dose of ATRA, and then again 3 h after dosing on Weeks 3 and 8 of treatment. (A) Average plasma levels (± SE) declined over time, n = 20; (B) individual plasma levels were observed to increase over time in six subjects (30%); (C) levels increased after the first dose but declined after Week 3 of treatment in five subjects (25%); (D) in nine subjects (45%), plasma levels were highest after the first dose and declined with time.

Serial Pulmonary Function

On average, most measurements of pulmonary function, including FEV₁, FEV₁/FVC, vital capacity (VC), FRC, DL_CO, and C_stat, remained stable over the 6 mo of the study without significant change during either the placebo or ATRA treatment phases (Table 3).

Changes in Lung Anatomy as Determined by HRCT

Before quantitative image analysis, serial scans from each subject were evaluated for reproducibility by comparing the position of regional parenchymal and thoracic landmarks on corresponding images. Excellent reproducibility was achieved on all three visits in 16 of 20 patients. In the remaining four subjects, good matching was documented pre- and post-ATRA, but landmarks differed slightly when compared with baseline. For the purpose of illustration, matched images from the upper lobe of one patient are presented in Figure 2A. In addition to demonstrating reproducibility, these images also depict the regional heterogeneity of disease. The extent of emphysema was determined in a blinded manner by computerized QIA using a density mask technique set to a threshold of 910 HU. Using this technique, the percentage of whole lung tissue involved with emphysema at baseline averaged 38.3% (± 16.86) with a range from 12.1% to 70.7% (Table 1). As expected for predominantly tobacco-related emphysema, the distribution of disease was heterogeneous, with the right lung 41.1 ± 17.2%) usually more involved than the left lung (34.0 = 16.7%), and the upper lobes (43.6 ± 18.9% on the right and 36.3 ± 17.8% of the left) generally more involved than the lower lobes (38.6 ± 15.4% on the right and 34.4 ± 15.6% on the left). When whole lung density masks were compared over time, no overall difference in the extent of emphysema was observed comparing before with after ATRA treatment (Figure 2B and Table 4).

View larger version (60K): [in this window] [in a new window]   Figure 2.   (A) Axial CT images from one patient demonstrating reproducible anatomic registration of parenchymal and thoracic structures at TLC for each of the three CT examinations. (B) Corresponding plots for this patient depict the percentage of right and left lung tissue involved with emphysema according to whole lung density mask analysis (emphysema defined as voxels registering less than 910 HU). Density mask analysis yielded stable and reproducible measures over time.

Effect on QOL Scores

The SGRQ was used as a subjective measure of respiratory-related QOL. No significant differences in total score or any of the section scores were observed.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The discovery that ATRA reverses structural and functional lung damage in an elastase-induced model of emphysema (5) generated tremendous interest within both the lay and scientific communities. That finding, in conjunction with prior studies linking retinoids to the processes of alveolar septation and tissue matrix remodeling (3, 4, 19), provided the scientific impetus for this pilot study. We set out to address several specific issues. First, could ATRA be administered safely to elderly patients with moderate to severe emphysema? Would an intermittent dosing regimen result in stable drug levels over this extended treatment period? Lastly, we asked whether 3 mo of retinoid therapy improved physiologic and/or anatomic changes as measured by pulmonary function test and chest CT scans. Our results suggest that retinoid therapy for human emphysema is feasible and provides the groundwork for future clinical investigations.

ATRA was administered at an oral dose of 50 mg/m²/d following general guidelines for the treatment of APL. On a weight-adjusted basis, this represents approximately twice the dose administered in the rat emphysema model (5). All patients received 3 mo of active treatment. Side effects were all mild and either self-limited or responded to standard over-the-counter remedies. Because hyperlipidemia has been consistently reported as a dose-limiting toxicity in previous clinical trials using ATRA (7, 9), all potential subjects with baseline hypercholesterolemia or hypertryglyceridemia were started on lipid-lowering agents under the direction of their treating physicians. Using this strategy, only one patient developed grade 2 hypertryglyceridemia, which resolved spontaneously. Another concern with prescribing retinoids is lung cancer risk. Previous large scale chemoprevention studies have demonstrated that the combination of active smoking and retinoids, but not retinoids alone, increases the risk for lung cancer (23, 24). As such, we limited our investigations to former smokers. Under the conditions studied, treatment with ATRA was well tolerated.

When administered on a daily basis, ATRA induces its own oxidative catabolism within 3-5 d of beginning therapy, and plasma ATRA levels decrease on average by 60-70% when given continuously for even 1 wk (25, 26). This autoinduction may be responsible for the retinoid resistance syndrome that occurs in some patients treated for APL. We attempted to limit retinoid resistance by employing an intermittent dosing regimen. To evaluate the effect of this regimen, we serially measured plasma ATRA levels at 3 h after the first dose and again 3 h after the first dose of medication on Weeks 3 and 8 of treatment. Although less quantitative than a complete area under the curve analysis, our results suggest significant variability among subjects. Approximately one-third of patients developed a 50% or greater reduction in their plasma ATRA level over time. As such, our intermittent dosing schedule was only partially effective in preserving blood levels. These results heighten interest in other retinoids, such as 13-cis-retinoic acid and 9-cis-retinoic acid, which do not induce their own metabolism. 9-cis-Retinoic acid was recently shown to be equally potent in promoting lung regeneration in the elastase-induced emphysema model (6).

Emphysema is a heterogeneous process that varies widely in pattern and distribution from one subject to another. In designing outcome measures for this study, we included standard PFTs, measurement of lung compliance and gas exchange, CT scans, and QOL questionnaires. PFTs act as global measures of lung function and are relatively insensitive to regional changespotentially missing early changes in lung remodeling. The reproducibility of PFTs is also complicated by variability in effort, technique, bronchospasm, anemia, and other factors (12, 13). Chest CT, on the other hand, focuses on anatomic changes, and has the potential to evaluate tissue changes at the local and regional level (27, 28). Therefore, CT may provide the ability to detect subtle, regional changes that are not identifiable by PFT. Finally, QOL scores were employed as a whole body measurement that integrates changes in lung function, cardiac function, or drug-related effects/toxicity that are not detected by other means (18).

The crossover design allowed all participants to be treated with ATRA and proved to be an important recruitment strategy. It also provided the opportunity to obtain serial measurements during both the placebo and active treatment phases in a blinded manner. As such, it allowed us to examine outcome measures for reproducibility, stability, and therapy-associated changes. Our subjects represented a cross section of individuals with moderate to severe emphysema. On average, their physiologic parameters did not change significantly during any study phase except for RV and TLC, which tended to increase over time in all phases of the study. No definitive treatment-associated changes in standard PFTs, lung compliance, or gas exchange were observed. Our basic CT measurements of TLC and density masking also failed to identify overall changes in response to ATRA. However, more sophisticated CT analysis employing quantitative imaging of tissue matrix features and time-attenuation plots at the regional level are still being explored. Although SGRQ scores were either stable or worsened during active treatment with ATRA, they were improved in 6 of 10 subjects when queried 3 mo later. This trend, which did not reach significance, might imply that benefit from ATRA does not manifest itself immediately after treatment, but rather as a delayed response. However, considering the pilot nature of this study and the lack of statistical power, caution is warranted before drawing any conclusions about the clinical utility, or lack of, resulting from this treatment.

In summary, our experience with this pilot study suggests that a 50 mg/m²/d dose of ATRA is well tolerated, with only mild side effects, in patients with advanced emphysema. Pretreatment of hyperlipidemia, a common baseline finding in these patients, might have reduced the incidence of this dose-limiting toxicity and could be used as a strategy to allow higher doses of, and longer treatment with ATRA to be considered in future studies. The use of an intermittent dosing schedule, 4 d of ATRA every week, reduced the incidence of autoinduced drug metabolism, but was still associated with declining ATRA levels over time. Retinoids with similar biological activity but different pharmacokinetics, such as 13-cis-retinoic acid and 9-cis-retinoic acid, might offer some therapeutic advantage in this respect. Finally, more sensitive measures of lung remodeling such as biomarkers for matrix metaloproteinases, elastin synthesis, or surfactant production might provide insight into tissue modulation that has not yet advanced to a level detectable by PFTs or CT. Considering the heterogeneity of the disease process, further investigation into quantitative CT imaging at the regional level is also warranted. Although our pilot study failed to demonstrate that a 3-mo course of ATRA reverses emphysema, our results support the safety and feasibility of continued investigations into the utility of retinoids as a clinical therapy for emphysema.