Soluble Tumor Necrosis Factor Receptor: Enbrel (Etanercept) for Subacute Pulmonary Dysfunction Following Allogeneic Stem Cell Transplantation

Abstract

Subacute lung disease, manifested as either obstructive (OLD) or restrictive (RLD) lung dysfunction, is a common complication following allogeneic stem cell transplantation. In each case, therapeutic options are limited, morbidity remains high, and long-term survival is poor. Between 2001 and 2008, 34 patients with noninfectious, obstructive (25) or RLD restrictive lung dysfunction (nine) received etanercept (Enbrel®, Amgen Inc.) 0.4 mg/kg/dose, subcutaneously, twice weekly, for 4 (group A) or 12 weeks (group B). Corticosteroids (if present at study entry) were kept constant for the initial 4 weeks of therapy and then tapered as tolerated. Thirty-one of 34 (91%) subjects were evaluable for response, and 10 (32%) met primary response criteria. There was no difference in response based on the duration of treatment (29% group A versus 35% group B; P =.99), the presence of RLD or OLD (33% versus 32%; P =.73), or the severity of pulmonary disease at study onset. Estimated 5-year overall survival rates following therapy were 61% (95% confidence interval, 46%–80%) for all subjects and 90% (95% confidence level, 73% –100%) for the 10 who met the primary response criteria. Five-year survival estimates for subjects treated with RLD was 44%, compared with 67% for those treated for OLD (P =.19). Etanercept was well tolerated, with no bacteremia or viremia observed. Pathogens were noted on posttherapy bronchoalveolar lavage in two cases. These data support the development of expanded clinical trials to study etanercept as a therapeutic agent for subacute lung injury after allogeneic stem cell transplantation.

INTRODUCTION

Noninfectious lung injury is a frequent and severe complication of allogeneic stem cell transplantation (SCT), both in the acute and subacute setting. In the acute setting, a diffuse interstitial process termed idiopathic pneumonia syndrome (IPS) may occur, mediated by the production of inflammatory cytokines and associated with high mortality rates (>50%) [1–3]. Subacute lung injury, on the other hand, typically presents in patients over 100 days posttransplantation and is associated with significant morbidity and mortality within the first 1 to 2 years following SCT. Subacute lung injury has been well described and may present as obstructive (OLD) or restrictive (RLD) lung dysfunction [4–11].

OLD is characterized by enhanced airflow resistance upon expiration, reflecting narrowing or destruction of smaller airways and terminal bronchioles. Commonly associated with the occurrence of chronic graft-versus-host disease (cGVHD), obtructive defects have been reported in 2% to 25% of allogeneic SCT recipients [4,10–12] Bronchiolitis obliterans remains the most common histopathology associated with OLD. The clinical course is variable, ranging from a gradual decline in lung function over several years to a rapid pulmonary deterioration over a few months. Responses to standard immuno-suppressive therapy have been limited, usually resulting in preservation of, instead of improvement in, existing lung function [13,14].

RLD is associated with reductions in forced vital capacity (FVC), total lung capacity (TLC), and carbon monoxide diffusion capacity (DLCO). Restrictive defects are more frequent than obstructive changes following allogeneic SCT, with an incidence of 20% to 45% reported [6,8,12,15]. In comparison to OLD, restrictive defects typically present earlier, are more frequent following conditioning regimens with total-body irradiation, and have been found in association with the development of both acute (aGVHD) and cGVHD [16,17]. Similar to OLD, therapeutic intervention often stabilizes, without significantly improving respiratory function.

Over the past several years, preclinical data generated in our laboratory using murine transplantation models indicate that two distinct, but interrelated pathways of immune-mediated lung injury may exist: one driven by soluble inflammatory cytokines and the other by host antigen-specific T cell effectors. Both pathways can target lung tissue, resulting in inflammation and pulmonary injury. Specifically, preclinical data have revealed a critical role for tumor necrosis factor-alpha (TNF-α) in the development of IPS after allogeneic SCT [18,19], and these laboratory insights have formed the foundation for completed and ongoing clinical trials [20].

In contrast to acute lung injury, the pathophysiology of subacute lung injury is less clearly defined. The development of OLD is characterized by bronchiolar leukocyte recruitment leading to fibrinous-obliteration of the small airways [10,21–23]. The mechanism of injury likely involves injury to the bronchiolar epithelium followed by an on-going inflammatory response and dysregulated repair [24]. The severity of the injury parallels the duration of this inflammatory response. Similarly, the pathogenesis of RLD also appears to involve a chronic inflammatory process in the lung interstitium, with interactions between cytokine, chemokine, and cellular effectors [24]. Bronchoalveolar lavage (BAL) fluid from patients with bronchiolitis obliterans syndrome (BOS) following lung allograft transplantation reveals elevations in interleukin (IL)-1ra, IL-8, transforming growth factor-beta (TGF-β), and monocyte chemotactic protein-1 (MCP-1), all of which have been implicated in other fibroproliferative processes [25–28]. Similarly, elevations of TNF-α and TGF-β have been shown to play a critical role in models of interstitial lung disease and fibrosis [28–30].

Based on these findings, we initiated a clinical trial employing a TNF-α binding agent, etanercept (Enbrel®, Amgen Inc., Thousand Oaks, CA) for the treatment of patients with noninfectious, subacute lung injury following allogeneic SCT. Etanercept is a soluble dimeric protein consisting of two soluble p75 TNF receptors fused to the Fc portion of a type I immunoglobulin molecule [31–34]. The agent is FDA approved for the treatment of rheumatoid arthritis [35–38] and has shown efficacy and safety in clinical trials for the treatment of aGVHD and cGVHD and IPS [20,39,40]. We have now examined its potential role in the treatment of subacute lung injury following allogeneic SCT.

MATERIALS AND METHODS

Patients

Study subjects were recruited from the Blood and Marrow Stem Cell Program at the University of Michigan Medical Center between 2001 and 2008, all subjects having received an allogeneic SCT at least 100 days before study entry. Eligible subjects were ≥6 years in age, with evidence of subacute, noninfectious lung injury (OLD or RLD) based on abnormalities observed on pulmonary function tests (PFTs). Patients were excluded if an active infection was noted at the time of study entry, based on cultures obtained from BAL fluid, blood, and urine. Patients requiring inotropic blood pressure support or those patients exhibiting clinical or echocardiographic signs of left ventricular dysfunction were also excluded. The study was approved by the University of Michigan institutional review board, and all subjects provided informed consent/assent before initiation of study therapy.

Study Design

The diagnosis of subacute noninfectious lung injury (Table 1) was based on the presence of the following three criteria: (1) abnormal pulmonary function tests, (2) clinical findings, and (3) absence of lower respiratory tract infection based on negative BAL fluid or lung biopsy cultures for infectious pathogens. Subjects with subacute lung injury were stratified as having either RLD or OLD based on the predominant respiratory pattern. Restrictive lung defects were defined by TLC and FVC ≤80% predicted, with a >15% decline from the pretransplantation values. Obstructive lung defects were defined by both the FEV₁ and FEV₁/FVC, with the requirement for a FEV₁ <80% predicted (and FEV₁/FVC <75% predicted), with >15% decline in FEV₁ from the pretransplantation value. Postbronchodilator values for FEV₁ were used if available. The date of diagnosis of subacute lung injury was determined by the date of the initial PFT that met the above criteria.

Table 1.

Subacute Lung Injury—Required Study Criteria

1. Clinical symptoms:

a. Cough, wheezing, or dyspnea on exertion or

b. Oxygen saturation #93% on room air.

2. Abnormal pulmonary function tests:

>15% decline from pretransplantation value, plus

a. Restrictive disease: FVC <80% predicted, OR

b. Obstructive disease: FEV1 <80% with FEV1/FVC <75% predicted

3. Absence of lower respiratory tract infection

a. Negative bronchoalveolar lavage or

b. Negative transbronchial or surgical lung biopsy (if performed).

The 2005 National Institutes of Health (NIH) consensus criteria for BOS were applied prospectively to those subjects with OLD enrolled after 2005 [41]. Pulmonary function parameters for OLD were quantified using the NIH-recommended Lung Function Scoring (LFS) system [41]. Lung function scores were determined individually for FEV₁ and DLCO, with ≥80% = 1, 70% to 79% = 2, 60% to 69% = 3, 50% to 59% = 4, 40% to 49% = 5, and <40% = 6, based on % predicted spirometric values. The FEV₁ scores were summated with the DLCO score, to determine an overall LFS score.

All eligible subjects were evaluated by BAL, high-resolution computed tomography, and complete PFTs before study therapy (Figure 1). Transbronchial biopsies were not required but were considered on a per-subject basis. Upon completion of study therapy, subjects underwent repeat evaluation, including BAL, PFTs, chest high-resolution computed tomography, and serologic studies. In addition, PFTs and serologic studies were repeated 28 days after study completion. Subjects were treated with etanercept for either 4 weeks (group A) or 12 weeks (group B). Group assignment was stratified by year of study entry. Group A consisted of subjects treated between 2001 and 2002, whereas group B subjects were treated from 2003 to 2008. Supportive care measures for both group A and B were similar, except for the introduction of voriconazole as antifungal prophylaxis in subjects treated after 2005.

Figure 1.

Study schema.

Bronchoalveolar lavage

Patients were required to undergo BAL before study enrollment and at the end of therapy. Assessment of the pre- and posttherapy BAL fluid included Gram stain; quantitative bacterial culture; viral cultures for respiratory syncytial virus, parainfluenza, adenovirus, influenza A and B, herpes simplex virus, and cytomegalovirus (CMV); stains for fungi; fungal and mycobacterial cultures; stains for acid-fast bacilli, fungi, and Pneumocystis jiroveci pneumonia; and polymerase chain reaction for Pneumocystis jiroveci pneumonia. BAL fluid samples were collected from right- and left-sided airways, collectively pooled, and then subdivided for the above studies. The pretherapy BAL was performed within 2 weeks before initiation of etanercept therapy, and the posttherapy BAL was performed within 1 week following completion of etanercept.

Pulmonary Function Tests

For group A subjects, PFTs were performed pretherapy, week 4 (end of therapy), and at week 8 (posttherapy). For group B subjects, PFTs were performed pretherapy, then every 4 weeks during study therapy, at week 12 (end of therapy), and week 20 (posttherapy). In all cases, measurements of FVC, FEV₁, TLC, and DLCO were obtained using a SensorMedics Vmax 229 and Sensor-Medics Autobox 6200 System (Sensor-Medics Corp., Yorba Linda, CA). Mean values for subjects were expressed as percent predicted for individuals of the same height and age. Pretherapy PFTs were performed within 2 weeks before the BAL, and posttherapy PFTs were performed within 1 week of the defined end point.

Serologic Testing

Complete blood counts, serum electrolytes, BUN, creatinine, and liver function tests were performed at study entry, every 4 weeks during therapy, and 4 weeks posttherapy.

Cytokine Analysis

Blood for inflammatory cytokine analysis was obtained at study entry and at the completion of therapy for the first cohort of patients (group A). Samples were collected in heparinized tubes, and transported to the University of Michigan Cancer Center Immunology Core Laboratory for processing. The plasma compartment was separated, divided, and stored at −80°C as per standard operating procedures, until the time of analysis. In addition, plasma samples were obtained for inflammatory cytokine analysis from control patients that had signed informed consent for such sample acquisition. Two subject groups served as controls for the plasma cytokine studies: (1) SCT recipients who had an “uncomplicated transplantation course” through day 100 (nine subjects), and (2) from healthy volunteers (seven subjects). Subjects with an “uncomplicated transplantation course” received a myeloablative conditioning regimen and subsequently exhibited no signs of aGVHD, hepatic veno-occlusive disease, significant infections, or pulmonary dysfunction through day 100 posthematopoietic cell transplantation and did not develop subsequent subacute lung injury. Plasma samples were obtained on or around day 100 posttransplantation in this group of control subjects.

Determination of Protein Levels by ELISA

Frozen plasma samples were thawed and analyzed in batches for a panel of inflammatory proteins including TNF-α, TNFRI, TNFRII, sCD-14, TGF-β (Bio-Source, Camarillo, CA), along with IL-6, IL-8, and MCP-1 (BD BioSciences-Pharmingen, San Diego, CA). ELISA assays were performed as per the manufacturer’s guidelines in the Immunology Core Laboratory at the University of Michigan Medical Center. All samples and controls were run in duplicate. Limits of detection for individual assays are as follows: TNF-α (1 pg/mL), TNFRI (50 pg/mL), TNFRII (100 pg/mL), sCD-14 (1 ng/mL), TGF-β (2 pg/mL), interferon-gamma (IFN-γ) (1 pg/mL), IL-8 (1 pg/mL), IL-6 (2.5 pg/mL), and MCP-1(1 pg/mL).

Etanercept Therapy

Etanercept therapy was initiated provided that baseline BAL fluid special stains were negative and BAL fluid cultures were negative for a minimum of 48 hours. Etanercept was reconstituted from a lyophilized powder and administered subcutaneously at a dose of 0.4 mg/kg (max dose = 25 mg), twice weekly (72 to 96 hours between doses), for a total of eight doses (group A) or 24 doses (group B). Dosing adjustments were not made based on renal or hepatic function. No premedication was administered. Subjects were observed for local reactions at the injection site within 30 minutes of a dose. Vital signs, including temperature, blood pressure, heart rate, and respiratory rate were recorded before, then at 15 and 30 minutes following a dose. Group A subjects were observed in our medical facility with every dose, whereas group B subjects were observed for their initial one to two doses. Etanercept therapy was discontinued and not reinstituted, if any one of the following criteria were subsequently met: (1) cultures, special stains, or polymerase chain reaction analysis of BAL fluid, blood, or biopsy specimens became positive for a potentially infectious organism; (2) the patient developed signs and symptoms consistent with the diagnosis of sepsis syndrome; or (3) the patient developed CMV viremia and was believed to have active CMV disease.

Concomitant Immunosuppressive Therapy

Subjects on corticosteroids at study entry were required to continue corticosteroids without dose alteration during the initial 4 weeks of etanercept therapy. Corticosteroids could be subsequently tapered as clinically indicated. Subjects requiring an increase in corticosteroid dosing during therapy were considered therapy failures and taken off study at that point. However, such patients continued to be assessed for toxicity. Dosing adjustments for tacrolimus or cyclosporine were allowed, based on measurement of serum drug levels. Other immunosuppressive agents were to continue throughout the study therapy, without dosing adjustment, unless clinically warranted.

Response and Outcome

Response assessments were based on improvement in pulmonary function parameters, pretherapy to posttherapy. FVC measurements were used to assess response in subjects with primary restrictive lung defects, and FEV₁ and FEV₁/FVC ratios were used to assess response in subjects with obstructive lung defects. Changes in DLCO measurements, pre- and posttherapy, were also examined for all subjects. In addition, interval response assessments at weeks 4 and 8 of therapy were determined for group B subjects. Response was defined as a ≥10% improvement in the absolute value for FEV₁ (for obstructive defects) or FVC (for restrictive defects). To be evaluable for response, subjects must have completed ≥50% of scheduled etanercept doses, a minimum of four doses for group A and 12 doses for group B subjects. The primary study endpoint was response at week 4 (group A) or week 12 (group B). Secondary response was determined by assessment at week 8 (group A) or week 20 (group B).

Statistical Analysis

Overall survival (OS) from the start of study therapy was estimated using Kaplan-Meier methods; statistical significance for differences in OS among patient characteristics (obstructive or restrictive lung defects, duration of etanercept therapy, and duration of lung injury before study entry) was assessed with a log-rank test. Statistical significance for differences in proportions among subject groups was assessed with a χ² test of association. Statistical comparisons of plasma protein data were completed using a Student’s t-test. Statistical significance was defined as a P value <.05.

RESULTS

Fifty-seven patients were evaluated for study therapy. Of these, 23 were deemed ineligible based on the presence of pathogens in their pretherapy BAL fluid (20 patients), radiographic findings (one patient), or relapsed disease (two patients). Thirty-four subjects, ranging from 8 to 65 years in age, met eligibility criteria and were treated (Table 2). The median time from transplantation to onset of study therapy was 1,099 days (mean, 1,062 days; range, 173–2,920 days). Subjects exhibited significant pulmonary dysfunction at study onset, with pretherapy pulmonary function testing shown in Table 3. Nine subjects met the criteria for restrictive pulmonary defects and 25 for obstructive defects at study entry. Subjects with OLD had a mean NIH LFS of 8 (median, 7.5; range, 2–12) at the time of enrollment. All subjects exhibited signs or symptoms of systemic cGVHD, with diffuse cutaneous (22), oral mucosal (25), ocular (22), and hepatic involvement (eight) noted. Twenty-three of 34 subjects (68%) were on corticosteroid therapy at the time of study entry, with dosing ranging from 0.03 to 1.9 mg/kg/day (mean, 0.4 mg/kg/day; median, 0.3 mg/kg/day). Concurrent immunosuppressive therapy included calcineurin inhibitors (tacrolimus or cyclosporine) alone (five), calcineurin inhibitors plus corticosteroids ± mycophenolate (22), mycophenolate ± corticosteroids (five), or sirolimus (two). Nine subjects were receiving extracorporeal photopheresis (ECP) at the time of study entry. Fourteen subjects were on antimicrobial prophylaxis (azithromycin or levofloxacin), and 22 subjects were on antifungal prophylaxis, including fluconazole (10), voriconazole (10), itraconazole (one), and posaconazole (one) at study entry.

Table 2.

Patient Demographics: Clinical Features at Study Entry

Demographic Factor	Group A (n = 14)	Group B (n = 20)	Overall (n = 34)
Age (years)
Mean	36	40	39
Median	32	46	43
Range	17–52	8–65	8–65
Gender
Male	8	12	20
Female	6	8	14
Diagnosis
AML/MDS/MF	6	10	16
ALL	2	1	3
CML	1	4	5
NHL/CLL	3	3	6
Myeloma	1	1	2
Nonmalignant	1	1	2
HLA match
HLA matched	13	18	31
Mismatch	1	2	3
Cell source
PSC	8	17	25
Marrow	6	3	9
Donor type
MRD	8	14	22
URD	6	6	12
Chronic GVHD present
Yes	14	20	34
No	0	0	0
Corticosteroidsa
At study entry	14	14	28
Median dose (mg/kg/day)	0.4	0.2	0.3
Lung injury pattern
Restrictive	3	6	9
Obstructive	11	14	25
NIH Lung Function Score (LFS)b
Median	7	8	7.5
Mean	8	8	8
Range	2–11	4–12	2–12
Supplemental O2 required	1	2	3

AML indicates acute myelogenous leukemia; MDS, myelodysplasia; MF, myelofibrosis; CML, chronic myelogenous leukemia; NHL, non-Hodgkin lymphoma; ALL, acute lymphoblastic leukemia.

^aMedian equivalent dose of prednisone.

^bLFS was determined for subjects with obstructive lung disease only.

Table 3.

Pulmonary Function Parameters: Pretherapy (Groups A and B)

	Group A (n = 14)	Group B (n = 20)	Overall (n = 34)
FEV1 (%)
Median	49	52	51
Mean	47	52	50
Range	25–73	29–79	25–79
FVC (%)
Median	61	63	62
Mean	61	62	62
Range	28–92	39–86	28–92
DLCOcorr (%)
Median	65	60	61
Mean	63	59	60
Range	15–102	29–96	15–102

Values expressed as percent predicted for corresponding height and weight (FVC and FEV₁), or contemporaneously measured hemoglobin (DLCOcorr).

Response and Survival

Response was defined as a ≥10% improvement in the absolute value for FEV₁ (for obstructive defects) or FVC (for restrictive defects). Thirty-one of 34 subjects (91%) were evaluable for response, with three subjects completing <50% of scheduled dosing. The primary response assessment was made upon completion of etanercept therapy at week 4 (group A) or week 12 (group B). Ten of 31 subjects (32%) met the primary response criteria: four of 14 (29%) in group A and six of 17 (35%) in group B; P=.99. For the 10 responders, the median response was 17.8% (mean, 19.1%; range, 11.0%–26.6%). When response was determined using the change in % predicted value, the median improvement was 8.0% (mean, 8.5%; range, 4%–16%) for these same 10 patients. DLCO responses (≥10% improvement) were seen in five of 31 (16%) evaluable patients.

There was no difference in response based on the pattern of lung dysfunction; three of nine patients (33%) with RLD met the primary response criteria, compared with seven of 22 (32%) with OLD; P = .73. There were no differences in response based on the severity of disease at study onset. Subjects with an FVC or FEV₁ < 50% at study entry met response criteria in five of 17 cases (29%) compared with responses in five of 14 (36%) subjects with an FVC or FEV₁ ≥50% at study entry; P = .79. Responses were not improved even if higher baseline values were selected for analysis. For example, five subjects exhibited an FVC or FEV₁ > 60% at study entry, with only two of the five (40%) meeting the response criteria. Responses were seen in seven of 18 (39%) subjects with subacute lung injury <24 months in duration, compared with three of 13 (23%) subjects with subacute lung injury ≥24 months in duration; P = .59. Age did not affect the likelihood of response. Four of 10 (40%) patients <30 years in age responded, compared with six of 21 (29%) patients ≥30 years in age (P = .82). Likewise, there was no difference in response based on stem cell source, with responses noted in four of nine (44%) patients receiving marrow grafts, compared with six of 22 (27%) patients receiving peripheral stem cells as their donor source (P = .60).

Ten subjects (31%) met the secondary response criteria, defined by their response at week 8 (group A) or week 20 (group B). Six (35%) group B subjects had responded by week 12, increasing to eight responses (47%) by week 20; P = 0.94. For both subject groups (A and B), we were unable to assess the durability of response at later time points, as additional modifications in immunosuppressive therapy had been allowed.

Twenty-one of the 34 (62%) subjects enrolled remain alive, 2.5 to 9.6 years status posttherapy. Estimated 5-year survival from study entry was 61% (95% confidence interval [CI], 46%–80%) for all subjects. Estimated 5-year survival rates for subjects treated on group A and group B therapy were 64% and 56%, respectively (P = .70) (Figure 2A). Likewise, 5-year survival estimates for subjects treated with RLD was 44%, compared with 67% for those treated for OLD (P = .19) (Figure 2B). Among the 31 patients who completed therapy, 5-year survival estimates were 90% (95% CI, 73%–100%) for 10 patients who responded to therapy, compared with 55% (95% CI, 37%–83%) for the 21 patients that failed to meet response criteria; P = .07 (Figure 2C).

Figure 2.

(A) OS by duration of etanercept therapy, group A (4 weeks) versus group B (12 weeks). Group A (—). Group B (- - -). (B) OS by pattern of lung injury, obstructive (—) versus restrictive (- - -) defects. (C) OS by response to therapy, responders (—) versus nonresponders (- - -).

Toxicity

Etanercept dosing was well tolerated, with therapy completed in all 14 patients in group A and 17 of 20 group B subjects (Table 4). No infusion-related reactions (fever, blood pressure changes) or local site injection reactions were reported following the administration of etanercept. Grade 3 to 4 infectious complications occurred in five subjects, including mucormycosis pneumonia (one), aspergillus pneumonia (one), acute sinusitis (one), osteomyelitis (one), and a grade 3 fever without an identified source (one). No episodes of bacteremia or viremia were noted during study therapy. Posttherapy BAL was performed in 28 of the 34 cases. Only twice were potential pathogens—Aspergillus species (one), mucormycosis (one)—identified on posttherapy BAL fluid cultures. Osteomyelitis was noted within the first week of therapy in one patient and was felt to be unlikely related to study treatment. Hematologic toxicity was rare, with one group B subject developing a transient grade 3 thrombocytopenia. Two toxic deaths occurred while on study therapy, from progressive pulmonary dysfunction (one) and mucormycosis pneumonia (one). Two subjects died from relapse, the relapses occurring within 4 weeks following initiation of study therapy.

Table 4.

Grade 3 to 4 Toxicities (CTC v.3.0)

	Incidence (%)
Infection	5 (14)
Electrolyte	4 (11)
Hyperglycemia	3 (8)
Renal/GU	2 (6)
Hepatic (ALT, T.Bili)	2 (6)
Pain	2 (6)
Hypertension	1 (3)
Gastrointestinal	1 (3)
Central nervous system	1 (3)
Thrombocytopenia	1 (3)

Corticosteroid Dosing

The study was not designed to evaluate responses in other organ sites. However, corticosteroid dosing at study onset and completion was analyzed. Eleven of 31 evaluable subjects (35%) had a ≥10% reduction in corticosteroid dosing during study therapy. Five of 14 (36%) group A subjects had reductions in corticosteroid dosing, compared with six of 17 (35%) evaluable group B subjects. No change in corticosteroid dosing was made in 17 patients, and three subjects (all in group A) required an increase in corticosteroids while receiving etanercept therapy.

Plasma Cytokine Levels

Plasma samples were collected at study entry from all group A subjects, with subsequent protein levels analyzed and compared with transplantation controls (allogeneic SCT recipients without complications) and from healthy (nontransplantation) controls as described in Materials and Methods. As shown in Figure 3, subjects with RLD or OLD showed increases in plasma levels of some, but not all proteins evaluated, including sTNFRI, sTNFRII, IL-8, and TGF-β. Posttherapy plasma levels were not appreciably changed, other than the expected increase in sTNFRII levels associated with systemic etanercept administration (data not shown). Plasma levels for IL-6 and TNF-α were at the limit of detection and not different among groups (data not shown).

Figure 3.

Plasma samples were collected, and protein levels were analyzed from all patients in cohort 1 at the time of study entry as described in Materials and Methods. In addition, control plasma samples were collected at day 100 posttransplantation from allogeneic hematopoietic cell transplantation recipients without complications (Control [BMT]) and from healthy (nontransplanted) individuals (Control [NL]). *P <.05; **P <.01.

Ineligible Patients

As previously noted, 20 patients were enrolled but deemed ineligible for study therapy because of the presence of pathogens on their prestudy BAL. In all 20 cases, the subjects had no clinical signs of an acute infectious pneumonitis, with absence of fever, pleuritic chest pain, or other acute symptomatology at the time of enrollment. Occult fungal infections were noted in 13 of the 20 patients, including Aspergillus species (10), mucormycosis (one), C. krusei (one), and Cryptococcus neoformans (one). Other occult infections included Gram-negative bacteria (five), and mycobacteria (three), with mycobacteria noted concurrently with Aspergillus in one case. All 20 patients were treated for the identified pathogen, with three deaths from infectious pulmonary complications noted within 2 months of initiating antimicrobial therapy. In the remaining 17 patients, PFTs were obtained 6 to 12 months following initiation of antimicrobial therapy, with the target PFT parameter (FEV₁ or FVC) improving in only five of 17 patients. The remaining 12 patients all had a decline in their target PFT parameter despite antimicrobial therapy (median change, −17.4%; range, −2.0% to −36.8%).

DISCUSSION

Our study reports the use of a soluble TNF binding agent, etanercept, in the management of 34 allogeneic SCT recipients with subacute lung injury. The response rates seen in our trial were promising, given that the degree of pulmonary dysfunction observed in subjects at the time of study entry was severe, with a mean LFS of 8, a median FEV₁ 51% in patients with OLD, and a median FVC 62% in subjects with RLD. There are few “prospective” trials reported in the treatment of subacute lung injury post-SCT, with the majority of trials either retrospective in nature, anecdotal case series, utilizing variable response definitions, or focusing on disease stabilization rather than improvement. The potential benefits of systemic corticosteroids, inhaled corticosteroids, immunosuppressive therapy, bronchodilators, antimicrobial prophylaxis (azithromycin), anti-TNF antibody therapy (infliximab), ECP, and leukotriene inhibitors have all been examined in small studies (<15 patients), with objective responses ranging from 8% to 100% [42–49]. Even the role of etanercept in treating subacute lung injury post-SCT has been previously described, with three of four treated patients showing improvements in DLCO or subjective clinical responses [40]. Given the strict methodology and response definitions used in our study, our 32% response rate may serve as a benchmark for future clinical trials in this group of patients.

OS rates were likewise promising in our study, with estimated 5-year survival rates of 61% for all subjects and 90% for those responding to therapy. By comparison, Dudek [42] noted a 5-year posttransplantation survival rate of 10% among patients with BOS, with a 79% 5-year survival rate (post-BOS) for those responding to front-line therapy and 13% for those failing. Clark et al [11,14] reported a 35% survival rate within 3 years of diagnosis of BOS, with long-term survival rates ranging from 0% to 86% noted in other reports.

A major finding of this report is that the use of etanercept in heavily immune-suppressed transplantation recipients was not associated with significant infectious or noninfectious complications. Overall, the toxicity of etanercept therapy was minimal. Historically, etanercept has been associated with infectious risks in patients with other inflammatory conditions, including Crohn’s disease, rheumatoid arthritis, and psoriaris, and has been contraindicated in patients with sepsis syndrome [50–53]. Reactivation of latent hepatitis B, hepatitis C, and tuberculosis infections has been seen with prolonged etanercept usage in these patient populations [54–57]. Despite the intensity of immune suppression concurrently used in our study population, secondary infections were extremely rare during or following etanercept therapy. No other major noninfectious complications were noted. Two patients relapsed early while on study, within 1 and 4 weeks of study entry. The impact of the etanercept on these two relapses is not clear. The potential for TNF inhibitors to negatively affect the graft-versus-leukemia (GVL) effect of the donor graft has been speculated with infliximab but not reported with etanercept [58]. Whereas anti-TNF monoclonal antibodies such as infliximab bind both soluble and membrane-bound TNF, etanercept binds strictly to soluble TNF, theoretically decreasing its potential impact on GVL activity.

As noted, both RLD and OLD may contribute to significant morbidity and late mortality in allogeneic SCT recipients [14,16,17,21,23,59]. In each scenario, collagen deposition and the development of fibrosis either in the interstitial (RLD) or peribronchiolar space (OLD) are believed to contribute to the patterns of lung dysfunction displayed on pulmonary function testing. It has been hypothesized that OLD and RLD after allogeneic SCT likely result from an initial insult to lung parenchyma, followed by an ongoing dysregulated reparative process involving the interplay between recruited immune effector cells and the resident cells of the pulmonary vascular endothelium and interstitium. To this end, we have previously proposed a triphasic model of chronic, non-infectious lung injury after allogeneic SCT that involves early evidence for T cell activation, leukocyte recruitment, inflammatory cytokine release, and ultimately, the deposition of collagen and the development of fibrosis [24]. In this model, an allogeneic response initiates injury to lung epithelial cells that is followed by a dysregulated reparative response and scarring of either the terminal airways as in OLD or in the case of RLD, the interstitial space.

Preliminary data generated in this study showed that patients with OLD and RLD had increased plasma levels of inflammatory markers including sTNFRI, sTNFRII, IL-8, and TGF-β. A more detailed biomarker analyses of plasma and BAL fluid collected from the entire study population is in progress and should help to validate our preliminary findings and further illuminate immunologic mechanisms that may be operative in the development of chronic pulmonary dysfunction after SCT. In general, the factors that regulate the progression from acute to chronic airway inflammation and the development of fibrosis remain to be determined, but experimental data suggest that TNF-α may be a common thread during this transition. Our group has previously reported that TNF-α is an important mediator of acute lung injury after allogeneic SCT [18, 19]. Moreover, TNF-α directly contributes to the development of interstitial fibrosis in several non-SCT models [60–62], and abrogation of TNF-α signaling either by antibody neutralization or by using mutant mice deficient in both TNF receptors significantly reduces lung fibrosis [30,63]. Additional evidence for a role of TNF-α in the transition from acute to chronic lung injury comes from studies using transgenic rodents with targeted overexpression of TNF-α in the lungs [61,62]. In these studies, early histopathology included a lymphocytic infiltrate similar to that seen in experimental IPS models, whereas the histologic changes associated with more prolonged exposure to TNF-α include collagen deposition and the development of fibrosis. From a clinical standpoint, Raghu and colleagues [64] recently completed an exploratory placebo-controlled trial using etanercept in 88 subjects with idiopathic pulmonary fibrosis. Overall, treatment with etanercept was well tolerated. Although there were no differences in the predefined endpoints, a decreased rate of disease progression was seen in several physiologic, functional, and quality-of-life endpoints among subjects receiving etanercept for idiopathic pulmonary fibrosis.

Recent biomarker studies in patients with cGVHD additionally support a hypothesis that Th1 and Th2 cytokine production may play a role in the pathogenesis of the disorder [65]. Increased levels of Th1 cytokines, including IFN-γ, TNF-α, IL-1, and IL-6 have been noted in skin and renal biopsies and in peripheral blood lymphocytes collected from patients with cGVHD [65–67]. In addition, elevations in plasma cytokine levels, including TNF-α, have been shown to precede the clinical development of cGVHD [65,68]. Elevated levels of macrophage migration inhibitory factor, an important regulator of TNF-α production, have also been noted in patients with cGVHD [69], and reductions in plasma levels of TNF-α have been reported following ECP therapy for cGVHD [70]. A recent Children’s Oncology Group study reported increased plasma levels of T and B cell activation markers, including sBAFF, anti-dsDNA antibody, sIL-2R, and sCD13 in patients with early- and late-onset cGVHD [71]. Finally, clinical studies have uncovered a possible link between variations in genes encoding for components of the innate immune system that exist within donor-recipient pairs and the development of air-flow obstruction and BOS following allogeneic SCT [72,73].

There are a few limitations to this study that warrant comment. One could argue that current NIH consensus criteria for BOS [41] were not used to determine study eligibility for all 25 patients with OLD. The definition for OLD in the current trial predated the NIH consensus criteria, with the majority of patients (69%) enrolling before the NIH consensus publication. However, when NIH consensus criteria were applied, 22 of 25 (88%) patients with OLD in our study still met the diagnostic criteria for BOS, based on PFT results, radiographic findings, and negative BAL cultures/stains. The remaining three subjects met three of the four diagnostic criteria for BOS, narrowly failing to achieve the requisite FEV₁/FVC ratio <0.7 by 0.01–0.03, respectively. Our study requirement for a negative pretherapy BAL additionally indicates how strict we were in selecting our patients. Second, although response rates and survival curves in our study population are encouraging, comparisons can only be made to previously published reports. This was no placebo-controlled study nor was there a contemporary group of patients with subacute lung injury (treated with standard immunosuppression) available at the University of Michigan for comparison.

It is interesting to note that although 57 patients were enrolled on study, 20 (35%) were ineligible to begin study therapy because of the identification of pathogens on pretherapy BAL fluid cultures. The high incidence of occult infections (primarily fungal) in this study population indicates the need for bronchoscopy as a routine management tool in the evaluation of allogeneic SCT recipients with subacute lung dysfunction. In the majority of cases, the underlying pulmonary function defect (obstructive or restrictive) persisted following completion of appropriate antimicrobial therapy.

In summary, the current study was a prospective, open-label trial for patients with late-onset (subacute) lung injury following allogeneic SCT. The use of etanercept therapy was well tolerated in this patient population, with few complications observed and objective responses seen in 32% of patients. Long-term survival following etanercept therapy was excellent, with an estimated 5-year OS of 67% for patients with obstructive pulmonary disease, and 90% for those responding to 4 or 12 weeks of therapy. The response rates and toxicity profile support the development of larger phase II or randomized phase III trials to investigate the role of etanercept therapy in patients with subacute lung injury postallogeneic SCT.