- Email: [email protected]
Iron overload and nitric oxide-derived oxidative stress following lung transplantation
Iron Overload and Nitric Oxide– Derived Oxidative Stress Following Lung Transplantation David Reid, FRACP,a Gregory Snell, FRACP,a Christopher Ward, PhD,b Raj Krishnaswamy, MSc,b Roger Ward, MPhil,b Ling Zheng, MD,b Trevor Williams, FRACP,a and Haydn Walters, DMa Background: Reactive oxygen species (ROS) may contribute to airway injury and the development of the bronchiolitis obliterans syndrome (BOS) following lung transplantation (LT). Chemically active iron released from ferritin stores and nitric oxide (NO)-derived radicals may add to the oxidative burden. Methods: We determined the concentrations of ferritin and the aqueous NO derivative nitrite (NO2⫺) within bronchoalveolar lavage fluid (BALF) of 14 stable LT recipients (ST) and 7 subjects with BOS and 21 normal controls. We also assessed the relationship between BALF ferritin and hemosiderin-laden macrophages (HLMs) using a hemosiderin score (HS) and determined BALF albumin concentration as a marker of microvascular leakage. Results: BALF ferritin concentrations and HSs were significantly elevated in LT recipients overall compared with normal controls (p ⬍ 0.05). BALF NO2⫺ levels were elevated in BOS subjects and STs compared with normal controls (p ⫽ 0.002 and p ⫽ 0.09, respectively), but there was no difference between transplant groups. BALF albumin concentrations were elevated in BOS patients compared with normal controls (p ⫽ 0.02) and ST (p ⫽ 0.05), but there was no difference between STs and controls. There was a significant relationship between BALF ferritin concentration and HS in LT recipients overall (r(s) ⫽ 0.7, p ⬍ 0.001). In BOS subjects, but not ST, BALF ferritin was significantly related to BALF albumin (r(s) ⫽ 0.8, p ⫽ 0.05) and there was a weak relationship with NO2⫺ concentration (r(s) ⫽ 0.6, p ⫽ 0.1). BALF NO2⫺ was strongly related to BAL %neutrophils in BOS subjects (r(s) ⫽ 0.9, p ⬍ 0.01), but there was no such relationship in STs. Conclusions: Our findings suggest that the allograft could be subject to significant irongenerated oxidative stress, which may be exacerbated by NO and neutrophil-derived ROS, particularly in BOS. Microvascular leakage may be a feature of established chronic rejection, which potentiates the iron overload and contributes to further airway damage and remodeling. J Heart Lung Transplant 2001;20:840–849.
I
mprovements in surgical techniques, donor organ preservation solutions and immunosuppression regimens have resulted in lung transplantation (LT)
becoming a successful treatment for a variety of end-stage pulmonary diseases. Long-term survival is mainly dependent upon the absence of chronic
From the aDepartment of Respiratory Medicine, Alfred Hospital, Melbourne, Victoria, Australia; and bMonash University Medical School, Melbourne, Victoria, Australia. Supported by NH & MRC, Australia. Submitted October 11, 2000; accepted January 18, 2001. Reprint requests: Dr. E. H. Walters, P.O. Box 315, Prahran 3181
Victoria, Australia. Telephone: ⫹613-9276-3476. Fax: ⫹6139276-3434. E-mail: [email protected] Copyright © 2001 by the International Society for Heart and Lung Transplantation. 1053-2498/01/$–see front matter PII S1053-2498(01)00282-0
840
The Journal of Heart and Lung Transplantation Volume 20, Number 8
rejection, which is characterized histopathologically by obliterative bronchiolitis (OB), the incidence of which has not substantially decreased over the past decade.1 Histologic confirmation of OB is not straightforward because of its patchy nature and the lack of sensitivity of trans-bronchial biopsies (TBB).2 Bronchiolitis obliterans syndrome (BOS) describes the irreversible, progressive airflow obstruction, which, even in the absence of histologic confirmation, but with exclusion of any other identifiable cause, has become synonymous clinically with chronic rejection.1,3 BOS remains resistant to most interventions, although early identification may allow augmented immunosuppression and an improved prognosis.4 Recent interest has focused on the role of oxidative stress following LT. It has been suggested that the generation of free radicals (FRs) and reactiveoxygen species (ROS), particularly by activated neutrophils, is relevant to the chronic inflammatory processes involved in the development of BOS.5 Chemically reactive iron and aqueous nitric oxide (NO) derivatives are potential candidates that may contribute to ROS generation.6 – 8 Consistent with this, an excess of iron and up-regulation of inducible NO synthase (iNOS) expression within lung allografts has recently been described.9 –12 One potential source of iron within the allograft is ferritin, a soluble iron-binding protein, which comprises the primary storage molecule of iron within human tissues. Under normal conditions, intracellular ferritin functions to protect the cell against iron-generated ROS; however, local intracellular ferritin synthesis can be up-regulated by a number of inflammatory mediators, which can also cleave iron from extracellular ferritin’s protein shell, allowing “free iron” to participate in FR generation.13–16 The alveolar macrophage (AM) attempts to protect against these potential iron-catalyzed oxidative insults by scavenging iron and sequestering it as chemically inert hemosiderin and ferritin.17 Hemosiderin-laden macrophages (HLMs) are often noted anecdotally in routine BAL and TBB specimens from LT recipients and are usually considered (although with little, if any evidence) to be secondary to bleeding after previous biopsy procedures.18 However, we hypothesize that the presence of numerous HLMs in bronchoalveolar lavage samples following LT is a surrogate marker of a more generalized iron load. Increased microvascular permeability secondary to airway and/or lung parenchymal inflammation may be another potential source of iron within the
Reid et al.
841
airway in LT recipients. Although microvascular leakage is not well described in LT, the inflammatory processes involved in liberating ferritin and iron from cellular sources (i.e., NO and tumor necrosis factor-␣ [TNF-␣] production) have also been shown to increase pulmonary permeability, and similar mechanisms could be operating within the allograft.19,20 BALF albumin is elevated in some subjects with non–lung-transplant-related pulmonary inflammation,21,22 and in this study we have determined its concentration as a marker of lung vascular leakage following LT. To test our hypothesis that the lung allograft is subject to significant iron and NO-generated oxidative stress, we measured the concentrations of ferritin and the stable NO derivative nitrite (NO2⫺) in bronchoalveolar lavage fluid (BALF) of ST and BOS subjects compared with healthy, non-smoking normal controls.23 We formally quantified HLMs from LT recipients and assessed their relationship to BALF ferritin. We also determined BALF albumin concentrations and differential cell counts to determine whether increased pulmonary vascular permeability and polymorphonuclear cells (PMNs) might be relevant to the disease processes that contribute to the oxidative load.
MATERIALS AND METHODS Subjects and Equipment BALF samples from 21 subjects who had undergone LT at the Alfred Hospital between March 1991 and October 1994 were studied (Tables I and II). Among the study population, presence of BOS was defined as a “fixed” 20% or greater fall in forced expiratory volume in 1 second (FEV1) from the best achieved post-transplant in the absence of any other identifiable cause. Stable transplant recipients were defined as those who were clinically well, had well-preserved lung function (i.e., within 20% of their best posttransplant FEV1) and no more than A1 grading for acute lung rejection on standard histologic grading of trans-bronchial lung biopsies.24 Bronchoscopy was performed as part of routine surveillance or for investigation of deteriorating lung function in the BOS group. No bronchoscopic procedure included in this study was performed within 4 weeks of the last, and the median time from the previous procedure was 87 days (range 28 to 912 days). The transplant population studied consisted of 14 STs (median age 34 years, range 20 to 46 years; 8 women; median time from transplant 178 days, range 65 to 1,128 days) and 7 subjects with BOS
842
Reid et al.
The Journal of Heart and Lung Transplantation August 2001
TABLE I Clinical characteristics of stable lung transplant recipients Subject
Age (years)
Gender
Original Diagnosis
Operation
Time From Tx (Days)
Time From TBB (Days)
FEV1 liters/second
FEV1% Max
1 2 3 4 5 6 7 8 9 10 11 12 13 14
46 41 41 32 33 22 34 33 46 40 20 38 22 48
M F F F F F F M F F M M M F
Sarcoid PPH E CF E CF E E B B E CF CF COPD
BSLT HLT HLT BSLT HLT BSLT HLT HLT BSLT BSLT HLT BSLT BSLT (R)SLT
84 125 534 65 427 101 178 1128 108 178 410 542 84 375
29 73 178 44 101 37 93 359 63 87 199 186 28 84
3.4 2.5 3.4 1.8 3.6 3.4 3.5 3.4 1.9 1.9 4.3 3.2 3.1 2.0
100 100 100 95 100 100 100 100 100 100 99 91 100 92
CF, cystic fibrosis; PPH, primary pulmonary hypertension; E, Eisenmenger’s syndrome; ␣ 1-AT, ␣ 1-anti-trypsin deficiency; B, bronchiectasis; COPD, chronic obstructive pulmonary disease; BSLT, bilateral sequential lung transplant; HLT, heart–lung transplant; SLT, single lung transplant; L/R, left/right; Tx, transplant; TBB, trans-bronchial biopsy; FEV1% max, forced expiratory volume at 1 sec as a percentage of best achieved result post-transplant.
(median age 25 years, range 19 to 50 years; 4 women; median time from transplant 1,080 days, range 270 to 1,440 days). All BALF aspirates were routinely evaluated for evidence of viral, bacterial or fungal infection by conventional microscopic and culture techniques and, if significant infection was thought to be present, the sample was excluded from study. Twenty-one healthy, non-smoking volunteers (median age 23 years, range 19 to 38 years) were recruited to provide control data for BALF indices. All control subjects had normal lung function and were not taking any regular medications. The study was approved by our institutional ethics committee and written informed consent was obtained from each subject.
Bronchoscopic Procedure Flexible fiber-optic bronchoscopy was performed using topical anesthetic (lignocaine 4% above the cords and 2% below in 2-ml aliquots, total dose ⱕ 24 mg) and intravenous midazolam as a sedative. The bronchoscope was passed trans-nasally and wedged in a sub-segmental bronchus of either the right middle lobe or the lingula segment of the left upper lobe. Three 60-ml aliquots of sterile, phosphatebuffered saline solution, pre-warmed to 37°C, were injected and aspirated immediately with moderate suction (⫺50 to ⫺100 mm Hg) after each individual administration. The BALF aspirated was immediately placed into siliconized glass vessels on ice.
TABLE II Clinical characteristics of lung transplant recipients with bronchiolitis obliterans syndrome Subject
Age (Years)
Gender
Original Diagnosis
Operation
Time From Tx (Days)
Time From TBB (Days)
FEV1 liters sec
FEV1% Max
15 16 17 18 19 20 21
25 24 19 20 42 30 50
M M M M F F F
CF CF CF CF ␣ 1-AT PPH COPD
BSLT BSLT HLT HLT (L)SLT BSLT (R)SLT
270 540 1080 1080 1440 395 1100
72 30 226 218 912 54 464
2.52 1.3 1.9 2.6 0.6 0.5 0.7
75 55 64 52 28 21 49
CF, cystic fibrosis; PPH, primary pulmonary hypertension; E, Eisenmenger’s syndrome; ␣ 1-AT, ␣ 1-anti-trypsin deficiency; B, bronchiectasis; COPD, chronic obstructive pulmonary disease; BSLT, bilateral sequential lung transplant; HLT, heart–lung transplant; SLT, single lung transplant; L/R, left/right; Tx, transplant; TBB, trans-bronchial biopsy; FEV1% max, forced expiratory volume at 1 sec as a percentage of best achieved result post-transplant.
The Journal of Heart and Lung Transplantation Volume 20, Number 8
Reid et al.
843
Cell Staining and Counting The volume of BAL return was recorded and the fluid filtered through a 200-m-pore-size stainlesssteel filter. Total cell count was performed on unprocessed BALF using a hemocytometer. BALF cytospins were prepared by centrifuging at 850 rpm for 10 minutes using a Shandon II cytocentrifuge and stored at ⫺80°C, so that subsequent counting could be done in standardized runs. BALF was sub-divided into 1.8-ml aliquots and stored at ⫺80°C. No part of the aspirate was discarded. Stored frozen cytospots were subsequently taken from the freezer and thawed at room temperature for 30 minutes, then aired with methanol for 10 minutes and immediately washed with distilled water. Cytospots were stained with Perl’s reagent for 20 minutes, washed and counterstained with nuclear-fast red for 10 minutes. Slides were then washed, dried and coverslipped. A hemosiderin score (HS) was calculated as described by Kahn.25 Briefly, 200 macrophages were examined on each slide and each cell was ranked for hemosiderin content using the following scale: 0 ⫽ no color; 1 ⫽ faint blue in one portion of cytoplasm; 2 ⫽ deep blue in a minor portion of the cell; 3 ⫽ deep blue in most areas of the cytoplasm; and 4 ⫽ deep blue throughout the cell. The total value for all cells was calculated and divided by 2 to obtain a score for an average of 100 cells. In addition, the simple percentage of cells staining positively was also recorded. Cell counting and scoring was performed by one experienced observer blinded to subject category.
BALF Ferritin Ferritin concentration in cell-free, unconcentrated BALF was determined by fluorescence immunoassay using an AXSYM analyzer and expressed as micrograms per liter.
BALF Albumin Albumin concentration was assessed using an autoanalyzer (Cobas Fara II, Roche Diagnostics, Basel, Switzerland) and expressed as milligrams per liter.
FIGURE 1 Hemosiderin score for stable lung
transplant recipients (ST), subjects with BOS and normal controls.
2.6 mol/liter NaOH was added to maximize fluorescence. Fluorescence of the final reaction product, 1(H)-naphotriazole (NAT), was determined using a spectrometer (Aminco-Bowman, Rochester, NY) and results expressed as the (nano)molar concentration of nitrite.
Statistical Methods Results are expressed as medians and ranges. The Kruskal–Wallis test was used to compare the three subject groups and the Mann–Whitney test was used to evaluate differences between specific groups. Spearman’s rank correlation was used to examine relationships between variables. Analysis was performed using the MINITAB for WINDOWS statistics program. A two-tailed p ⬍ 0.05 was considered significant.
BALF Nitrite (DAN Assay)
RESULTS Hemosiderin Score (HS) and BAL Cell Differentials
BALF NO2⫺ concentrations were determined by a fluorometric assay using 2,3-diaminonaphthalene (DAN; Molecular Probes, Eugene, OR). One hundred microliters of DAN solution was added to each 400 l of BAL sample. After thorough mixing the sample was incubated in the dark for 10 minutes and
BAL cells from 19 of the 21 transplant subjects and 3 of the 11 controls examined stained positively for hemosiderin within macrophages (Figure 1 and Table III). HS was significantly higher in both ST (median 10.5, range 0 to 187) and BOS subjects (median 6.5, range 0 to 27) compared with normal
844
Reid et al.
The Journal of Heart and Lung Transplantation August 2001
TABLE III Bronchoalveolar lavage findings in lung transplant recipients and normal subjects
ST BOS Normal controls
HS
BALF Neutrophils %
BALF Ferritin (g/liter)
BALF Albumin (mg/liter)
BALF Nitrite (nmol/liter)
10.5a (0–187) 6.5b (0–27) 0 (0–2)
4.0c (0.6–17.7) 46.4d (15.8–87.4) 1.2 (1.0–3.3)
48.8e (10–950) 126f (1–205) 2 (0–16)
33.8 (7–328) 57.9g (34.1–309) 30.2 (17.8–63.4)
440h (256–1194) 537i (443–1044) 337 (256–531)
BALF, bronchoalveolar lavage fluid; BOS, bronchiolitis obliterans syndrome; HS, hemosiderin score; ST, stable transplants. Hemosiderin score aST and bBOS vs normal, p ⫽ 0.001 and p ⫽ 0.01, respectively. BALF neutrophils %: cST and dBOS vs normal, p ⫽ 0.008 and p ⬍ 0.001, respectively (BOS vs ST. p ⬍ 0.001). BALF ferritin: eST and fBOS vs normal, p ⬍ 0.001 and p ⫽ 0.03, respectively. BALF albumin: g BOS vs normal and BOS vs ST, p ⫽ 0.02 and p ⫽ 0.05, respectively. BALF nitrite: hST and iBOS v normal, p ⫽ 0.09 and p ⫽ 0.002, respectively.
controls (median 0, range 0 to 2; p ⫽ 0.001 and p ⫽ 0.01, respectively). The percentage of cells positive for hemosiderin was also significantly higher in ST (median 7.5%, range 0% to 74.5%) and BOS subjects (median 5.0%, range 0% to 10.0%) than in normal controls (median 0%, range 0% to 1.5%; p ⬍ 0.001). There was no significant difference in HS or percentage of positive cells between LT recipient groups, that is, with or without BOS (p ⫽ 0.3). BAL neutrophils were significantly elevated in both ST (median 4.0%, range 0.6% to 17.0%) and BOS subjects (median 46.4%, range 15.8% to 87.4%), compared with normal controls (median 1.2%, range 1.0% to 3.3%; p ⫽ 0.008 and p ⬍ 0.001, respectively), and there was a significant difference between the two LT recipient groups (p ⬍ 0.001).
BALF Nitrite BALF NO2⫺ concentrations were significantly higher in LT recipients overall (median 475 nmol/ liter, range 256 to 1,194 nmol/liter) compared with normal controls (median 337 nmol/liter, range 256 to 531 nmol/liter; p ⬍ 0.05), but there was no significant difference between LTR groups (Figure 4 and Table III).
Correlations There was a significant correlation between BALF ferritin concentration and HS (Figure 5) in LTR
BALF Ferritin BALF ferritin concentrations were significantly higher in all LT recipients overall (median 49.0 g/liter, range 1 to 950 g/liter) compared with normal controls (median 2 g/liter, range 0 to 16 g/liter, p ⬍ 0.01), but there was no significant difference between ST (median 48.8 g/liter, range 10 to 950 g/liter) and BOS subjects (median 126 g/liter, range 1 to 205 g/liter, p ⫽ 0.9) (Figure 2 and Table III).
BALF Albumin BALF albumin levels were significantly higher in BOS subjects (median 57.9 g/liter, range 34.1 to 309 g/liter) compared with normal controls (median 30.2 g/liter, range 17.8 to 63.4 g/liter; p ⫽ 0.02) and there was a strong trend toward higher levels compared with ST subjects (median 33.8 g/liter, range 7.0 to 328 g/liter; p ⫽ 0.05) (Figure 3 and Table III). However, there was no difference in BALF albumin between ST subjects and normal controls (p ⫽ 0.8).
FIGURE 2 BALF ferritin concentrations in stable lung transplant recipients (ST), subjects with BOS and normal controls.
The Journal of Heart and Lung Transplantation Volume 20, Number 8
Reid et al.
845
FIGURE 3 BALF albumin concentrations in stable
lung transplant recipients (ST), subjects with BOS and normal controls.
overall (r(s) ⫽ 0.7, p ⬍ .001). This relationship was much stronger in ST (r(s) ⫽ 0.7, p ⬍ 0.005) than BOS subjects (r(s) ⫽ 0.4, p ⫽ 0.3), although this probably reflects the small sample size and lack of statistical power in the latter group. Similarly, the relationship between BALF ferritin concentration and %hemosiderin-positive AM was stronger in ST (r(s) ⫽ 0.8, p ⬍ 0.005) than BOS subjects (r(s) ⫽ 0.6, p ⫽ 0.2), although as a group a significant relationship remained in LT recipients overall (r(s) ⫽ 0.7, p ⬍ 0.001). BALF ferritin and albumin concentrations were significantly related in BOS subjects (r(s) ⫽ 0.8, p ⫽ 0.05) despite the small numbers, but there was no such relationship in ST subjects. Similarly, there was a weak relationship between BALF ferritin and nitrite concentrations in BOS (r(s) ⫽ 0.6, p ⫽ 0.1), but not in ST. Furthermore, there was a strong correlation between BALF NO2⫺ concentration and BAL %neutrophils in BOS subjects (r(s) ⫽ 0.9, p ⬍ 0.001), but not in ST subjects.
Clinical Outcomes There was no relationship between HS and %hemosiderin-positive cells or BALF ferritin concentration and time from transplant or previous biopsy procedure. All 13 ST cases remained well during the initial 6 months following the study procedures, but 5 sub-
FIGURE 4 BALF nitrite concentrations in stable lung transplant recipients (ST), subjects with BOS and normal controls.
sequently developed BOS during long-term follow-up (median time from bronchoscopy 44 months, range 34 to 70 months), and 3 of these have died. Four of the 7 patients with BOS died during followup. There was no obvious relationship between HS and/or BALF solute indices at the time of this reported sampling and subsequent outcomes in either the ST or BOS groups.
DISCUSSION The current study demonstrates a significant increase in BALF ferritin concentration and HLM numbers in the lower respiratory tract of lung transplant recipients (LTRs). We believe these findings are very unlikely to just be the sequelae of previous biopsy procedures. In addition, we found elevated levels of nitrite in BALF of LTRs, suggesting the allograft may be subject to significant iron and NO-derived free-radical damage. Furthermore, we confirmed previous observations that BAL %neutrophils are increased following LT, even in apparently stable subjects, which suggests that neutrophil-mediated airway inflammation may persist despite apparent well-being and intense immunosuppression and could be relevant in the pathogenesis of chronic rejection. Subjects with BOS had significantly higher levels of BALF albumin com-
846
Reid et al.
The Journal of Heart and Lung Transplantation August 2001
FIGURE 5 Relationship between BALF ferritin concentration and hemosiderin score in all lung transplant recipients (i.e., ST and BOS) as a group.
pared with normal controls and there was a strong trend for higher levels compared with ST. In addition, there was a significant relationship between BALF albumin and ferritin levels in BOS, which may indicate a microvascular leak syndrome that could contribute to the presence of ferritin and HLM within the airway. Our findings add to the literature concerning the potential role of iron overload and oxidative stress following LT. A recent study by Baz et al6 demonstrated elevated levels of total iron in the putative epithelial lining fluid (ELF) of lung allografts as well as an increase in tissue iron deposition in a small number of stable LTRs. We confirmed these observations in a larger population of patients and demonstrated similar findings in subjects with established BOS. An excess of chemically available iron within the lower respiratory tract may promote the catalysis of toxic hydroxyl radicals, which could contribute to the pathogenesis of cell injury and airway inflammation following LT.6,13,26 Also, the fibrogenic influence of iron is well recognized in overload syndromes such as hemochromatosis and idiopathic pulmonary hemosiderosis, and an increase in lung iron may be part of a final common pathway that ultimately leads to growth factor gen-
eration, promotion of collagen gene transcription and airway fibrosis and remodeling.26 Under normal conditions, iron within the lower respiratory tract is present in extremely low concentrations, bound to ferritin and other iron-binding proteins that prevent hydroxyl radical formation and limit the availability to iron-requiring microorganisms.26 However, the lung allograft may be subject to a number of inflammatory stimuli that can disrupt iron homeostasis and increase its chemical availability. Intracellular ferritin mRNA transcription and production can be up-regulated by TNF-␣, neutrophil-derived hydrogen peroxide and NO, all of which have been implicated in the processes of reperfusion injury, acute rejection and frequent infections following LT.12–15,27,28 A generalized increase in ferritin synthesis and storage could function as a reservoir of potentially reactive iron if intracellular stores are released secondary to cell injury or death. Extracellular ferritin will release its core of iron atoms in the presence of PMN and NO-derived radicals and our finding of increased PMN numbers and nitrite concentrations, in the setting of elevated BALF ferritin, suggests the allograft is likely to be subject to significant iron toxicity.16,29,30
The Journal of Heart and Lung Transplantation Volume 20, Number 8
To protect against these oxidative stresses the lung has evolved several defense mechanisms, which prominently involve the AM. These produce numerous anti-oxidant compounds, such as metallothionein and glutathione, which are detectable within ELF following LT and sequester iron intracellularly as chemically inert hemosiderin and ferritin.17,31,32 HLMs are seen following inhalation of iron-containing dusts, after alveolar hemorrhage and in cigarette smokers, but they have also been described in pulmonary fibrosis, suggesting they may be involved in the broader processes of lung inflammation and remodeling.17,18,33 Paradoxically, increased uptake and storage of iron within AMs may, under certain circumstances, enhance iron toxicity, particularly if intracellular ferritin synthesis lags behind the rate of iron uptake.34 This may be of particular relevance in LT where the ability of the AMs to leave the allograft is impaired because of diminished mucociliary clearance and lymphatic drainage. Thus, iron sequestration in lung allografts by AMs may be a “double-edged sword” and the relationship between BALF ferritin and HLMs in our study could be especially harmful if it reflects continuous cycling of ferritin-bound iron that may well have been released by injured bronchial epithelial cells and senescent oxidant-damaged AMs. Indeed, we have current data, as yet unpublished, from FACS analysis, which suggests that AMs are functionally impaired in LTR, and iron toxicity may be a contributing factor.35 The unknown lifespan of an AMs in lung allografts raises the potential criticism that the presence of HLMs within BALF in our study is simply representative of accumulated iron secondary to hemorrhage from previous biopsy procedures. We believe this to be unlikely for a number of reasons. First, BALF was obtained deliberately from either the right middle lobe or lingula segment of the left upper lobe specifically to avoid the sites ordinarily used for biopsy purposes at our center. In addition, despite the uncertainty concerning the time required for HLMs to leave the airway their lifespan probably remains limited and, under normal conditions, could be expected to disappear within 2 weeks of hemorrhage.18,36 In our study no TBB was undertaken within 28 days of that most recently performed procedure and there was absolutely no relationship between time from transplant or previous biopsy and BALF ferritin concentration or HS, which suggests that bleeding from biopsy procedures is not relevant. It is also
Reid et al.
847
highly improbable that the elevated levels of BALF ferritin in LTRs (median 49 ⫻ 10⫺6 g/liter) in our study are related to previous alveolar hemorrhage. This is because red blood cells (RBCs) contain extremely low concentrations of ferritin (⬍10⫺19 g/liter) and, according to our calculations, it would require a blood loss of at least 10 liters to account for the ferritin detected if this were the only source; that is, 10 (liters) ⫻ 5 ⫻ 10⫺12 (RBCs per liter of blood) ⫻ 10⫺19 (RBC ferritin content, in micrograms per liter) ⫽ 50 ⫻ 10⫺6 g/liter. Furthermore, although vascular leakage per se may be responsible for some of the ferritin detected, plasma exudation either as a consequence of airway inflammation or the lavage procedure itself is unlikely to be a significant contributor. A crude estimate of the “true” ELF concentration of ferritin can be calculated if one assumes a lavage dilution factor of at least 100, which equates to a median ELF ferritin concentration of 49 ⫻ 10⫺4 g/liter, which clearly greatly exceeds normal plasma levels (15 to 300 ⫻ 10⫺6 g/liter). In this study we were unable to demonstrate a significant increase in BALF albumin in LTRs as a whole compared with controls, but the elevated BALF albumin levels in BOS subjects probably does represent an increase in microvascular leakage. The latter could be occurring as a consequence of NO acting upon the vascular endothelium, although the absence of a direct relationship between BALF nitrite and albumin concentrations suggests other factors are also operative. Furthermore, although the trend toward a relationship between BALF ferritin and albumin in LTRs overall and a significant relationship in BOS subjects suggests vascular leakage may be responsible for some of the ferritin seen, it is perhaps not too surprising in our study that an isolated BALF albumin correlated poorly with HS, because of the difference in temporal history of their relative accumulations in the allograft. In conclusion, evidence is accumulating to support the hypothesis that the lung allograft is subject to significant oxidative stress mediated by neutrophil production of ROS and probably exacerbated by NOderived radicals, and especially the presence of an excess of tissue iron. The allograft may have an impaired ability to combat these insults because of antioxidant depletion and cell damage continuing unabated, with the eventual development of chronic rejection, airway remodeling and ultimately organ failure.31,32,37–39 The presence of ferritin and HLMs in BALF does not merely represent hemorrhage after
848
Reid et al.
recent lung biopsies, but rather reflects a generalized iron overload that may be of etiologic significance in the evolution of chronic rejection. Therapeutic interventions might involve dietary supplementation with anti-oxidants or iron chelation to reduce the oxidative burden and animal studies of acute rejection in cardiac and renal transplantation have demonstrated the benefits of such potential treatments.41,42 Prospective studies now need to be undertaken to address these issues and examine the relationship between allograft iron deposition, oxidative stress and anti-oxidant depletion and to relate these interactions to the development of chronic rejection. REFERENCES 1. Valentine VG, Robbins RC, Berry GJ, Patel HR, Reichenspurner H, Reitz BA, Theodore J. Actuarial survival of heart–lung and bilateral sequential lung transplant recipients with obliterative bronchiolitis. J Heart Lung Transplant 1996;15:371– 83. 2. Yousem SA, Paradis IL, Dauber, JH, et al. Efficacy of transbronchial lung biopsy in the diagnosis of bronchiolitis obliterans in heart–lung transplant recipients. Transplantation 1989;47:893–5. 3. Yousem SA, Suncan SR, Ohori P, Sonmez-Alpan E. Architectural remodelling of lung allografts in acute and chronic rejection. Arch Pathol Lab Med 1992;116:1175– 80. 4. Snell GI, Esmore DS, Williams TJ. Cytolytic therapy for the bronchiolitis obliterans syndrome complicating lung transplantation. Chest 1996;109:874 – 8. 5. Riise, GC, Williams A, Kjellstrom C, Schersten H, Andersson BA, Kelly FJ. Bronchiolitis obliterans syndrome in lung transplant recipients is associated with increased neutrophil activity and decreased anti-oxidant status in the lung. Eur Resp J 1998;12:82– 8. 6. Britton RS, Tavill AS, Bacon BR. Mechanisms of iron toxicity. In: Brock JH, Halliday JW, Pippard MJ, Powell LW, eds. Iron metabolism in health and disease. London: Saunders; 1994:311–51. 7. Freeman B. Free radical chemistry of nitric oxide. Looking at the dark side. Chest 1994;105(Suppl):79S– 84S. 8. Saleh D, Ernst P, Lim S, Barnes PJ, Giaid A. Increased formation of the potent oxidant peroxynitrite in the airways of asthmatic patients is associated with induction of nitric oxide synthase: effect of inhaled glucocorticoid. FASEB J 1998;12:929 –37. 9. Baz MA, Ghoi AJ, Roggli VL, Tapson VF, Piantadosi C. Iron accumulation in lung allografts after transplantation. Chest 1997;112:435–9. 10. Giaid A, Corris PA, Chikhani N, et al. Expression of nitric oxide synthase in lung transplant recipients with bronchiolitis obliterans. Eur Respir J 1995;8(Suppl):550S. 11. Gabbay E, Walters EH, Orsida B, et al. In stable lung transplant recipients, exhaled nitric oxide levels positively correlate with airway neutrophilia and bronchial epithelial iNOS. Am Rev Respir Crit Care Med 1999;160:2093–9. 12. Devlin J, Palmer RM, Gonde CE, et al. Nitric oxide generation: a predictive parameter of acute allograft rejection. Transplantation 1996;61:745–9.
The Journal of Heart and Lung Transplantation August 2001 13. Gutteridge JMC, Mumby S, Quinlan GJ, Chung KF, Evans TF. Pro-oxidant iron is present in human pulmonary epithelial lining fluid; implications for oxidative stress in the lung. Biochem Biophys Res Commun 1996;220:1024 –7. 14. Miller CC, Miller SC, Torti SV, Tsuji Y, Torti FM. Ironindependent induction of ferritin H-chain by tumor necrosis factor. Proc Natl Acad Sci USA 1991;88:4946 –50. 15. Kwak EL, Larochelle DA, Beaumont C, et al. Role for NF- in the regulation of ferritin H by tumour necrosis factor. J Biol Chem 1995;270:15285–93. 16. Reif DW. Ferritin as a source of iron for oxidative damage. Free Rad Biol Med 1992;12:417–27. 17. Corhay JL, Weber G, Bury T, et al. Iron content in human alveolar macrophages. Eur Respir J 1992;5:804 –9. 18. Grebski E, Hess T, Hold G, Speich R, Russi E. Diagnostic value of haemosiderin laden macrophages in bronchoalveolar lavage. Chest 1992;102:1794 –9. 19. Bernareggi M, Mitchell JA, Barnes PJ, Belvisi MG. Dual action of nitric oxide on airway plasma leakage. Am J Crit Care Med 1997;155:869 –74. 20. Stephens KE, Ishizaka A, Larrick JW, Raffin TA. Tumour necrosis factor causes increased pulmonary permeability and edema. Am Rev Respir Dis 1987;137:1364 –70. 21. Jones KP, Edwards JH, Reynolds SP, Peters TJ, Davies BH. A comparison of albumin and urea as reference markers in bronchoalveolar lavage fluid from patients with interstitial lung disease. Eur Resp J 1990;3:152– 6. 22. Fick RB, Metzger JW, Hal RB, et al. Increased bronchovascular permeability after allergen challenge in sensitive asthmatics. J Appl Physiol 1987;63:1147–55. 23. Tayeh MA, Marletta MA. Macrophage oxidation of Larginine to nitric oxide, nitrite, and nitrate. J Biol Chem 1989;33:19654 – 8. 24. Yousem SA, Novick RJ, Breen TJ, et al. A working formulation for the standardisation of nomenclature in the diagnosis of heart and lung rejection: lung rejection study group. J Heart Transplant 1990;9:593– 601. 25. Kahn FW, Jones JM, England DM. Diagnosis of pulmonary haemorrhage in the immunocompromised host. Am Rev Respir Dis 1987;136:155– 60. 26. Mateos F, Brock JH, Perez-Arellano JL. Iron metabolism in the lower respiratory tract. Thorax 1998;53:594 – 600. 27. Silkoff PE, Caramori M, Tremblay L, et al. Exhaled nitric oxide in human lung transplantation; a non-invasive marker of acute rejection. Am J Respir Crit Care Med 1998;157:1822– 8. 28. Gabbay E, Walters EH, Orsida B, et al. Post lung transplant bronchiolitis obliterans syndrome (BOS) is characterized by increased exhaled nitric oxide levels and epithelial iNOS. Am J Respir Crit Care Med 2000;162:2182–7. 29. Pantopoulos K, Weiss G, Hentze MW. Nitric oxide and oxidative stress control mammalian iron metabolism by different pathways. Mol Cell Biol 1996;16:3781– 8. 30. Reif DW, Simmons RD. Nitric oxide mediates iron release from ferritin. Arch Biochem Biophys 1990;283:537– 41 31. Courtade M, Carrera G, Paternain JL, et al. Metallothioein expression in human lung and its varying levels after lung transplantation Chest 1998;113:371– 8. 32. Baz MA, Tapson VF, Roggli VL, Van Trigt P, Piantadosi CA. Glutathione depletion in epithelial lining fluid of lung allograft patients. Am J Respir Crit Care Med 1996;153: 742– 6. 33. Rich EA. Iron metabolism by smokers’ alveolar macro-
The Journal of Heart and Lung Transplantation Volume 20, Number 8
34.
35.
36.
37.
38.
phages: protective or problematic in the lung? J Lab Clin Med 1988;111:598 –9. Wesselius LJ, Williams WL, Bailey K, et al. Iron uptake promotes hyperoxic injury to alveolar macrophages. Am Rev Respir Crit Care Med 1999;159:100 – 6. Ward C, Whitford H, Snell GI, et al. Bronchoalveolar lavage macrophage and lymphocyte phenotypes in lung transplant recipients. J Heart Lung Transplant. In press. Sherman MJ, Winnie G, Thomassen MJ, Abdul-Karim FW, Boat TF. Time course of haemosiderin formation and clearance by human pulmonary macrophages. Chest 1984;86:409 –11. Shivakumar BR, Kolluri SV, Ravindranath V. Glutathione homeostasis in brain during reperfusion following bilateral carotid artery occlusion in the rat. Mol Cell Biochem 1992; 111:125–9. Riise GC, Williams A, Kjellstrom C, Schersten H, Andersson BA, Kelly FJ. Bronchiolitis obliterans syndrome in lung
Reid et al.
39.
40.
41.
42.
849
transplant recipients is associated with increased neutrophil activity and decreased anti-oxidant status in the lung. Eur Respir J 1998;12:82– 8. Kobayashi H, Nonami T, Kurokawa S, et al. Changes in the glutathione redox system during ischaemia and reperfusion in rat liver. Scand J Gastroenterol 1992;27:711– 6. Martelius T, Scholz M, Krogerus L, et al. Antiviral and immunomodulatory effects of desferrioxamine in cytomegalovirus-infected rat liver allografts with rejection. Transplantation 1999;68:1753– 61. Slakey DP, Roza AM, Pieper GM, et al. Delayed cardiac allograft due to combined cyclosporine and antioxidant therapy. Transplantation 1993;56:1305–9. Wang C, Salahudeen AK. Cyclosporine nephrotoxcity: attenuation by an antioxidant inhibitor of lipid peroxidation in vitro and in vivo. Transplantation 1994;58:940 – 6. the previous procedure was 87 days (range 28 to 912 days).
I
mprovements in surgical techniques, donor organ preservation solutions and immunosuppression regimens have resulted in lung transplantation (LT)
becoming a successful treatment for a variety of end-stage pulmonary diseases. Long-term survival is mainly dependent upon the absence of chronic
From the aDepartment of Respiratory Medicine, Alfred Hospital, Melbourne, Victoria, Australia; and bMonash University Medical School, Melbourne, Victoria, Australia. Supported by NH & MRC, Australia. Submitted October 11, 2000; accepted January 18, 2001. Reprint requests: Dr. E. H. Walters, P.O. Box 315, Prahran 3181
Victoria, Australia. Telephone: ⫹613-9276-3476. Fax: ⫹6139276-3434. E-mail: [email protected] Copyright © 2001 by the International Society for Heart and Lung Transplantation. 1053-2498/01/$–see front matter PII S1053-2498(01)00282-0
840
The Journal of Heart and Lung Transplantation Volume 20, Number 8
rejection, which is characterized histopathologically by obliterative bronchiolitis (OB), the incidence of which has not substantially decreased over the past decade.1 Histologic confirmation of OB is not straightforward because of its patchy nature and the lack of sensitivity of trans-bronchial biopsies (TBB).2 Bronchiolitis obliterans syndrome (BOS) describes the irreversible, progressive airflow obstruction, which, even in the absence of histologic confirmation, but with exclusion of any other identifiable cause, has become synonymous clinically with chronic rejection.1,3 BOS remains resistant to most interventions, although early identification may allow augmented immunosuppression and an improved prognosis.4 Recent interest has focused on the role of oxidative stress following LT. It has been suggested that the generation of free radicals (FRs) and reactiveoxygen species (ROS), particularly by activated neutrophils, is relevant to the chronic inflammatory processes involved in the development of BOS.5 Chemically reactive iron and aqueous nitric oxide (NO) derivatives are potential candidates that may contribute to ROS generation.6 – 8 Consistent with this, an excess of iron and up-regulation of inducible NO synthase (iNOS) expression within lung allografts has recently been described.9 –12 One potential source of iron within the allograft is ferritin, a soluble iron-binding protein, which comprises the primary storage molecule of iron within human tissues. Under normal conditions, intracellular ferritin functions to protect the cell against iron-generated ROS; however, local intracellular ferritin synthesis can be up-regulated by a number of inflammatory mediators, which can also cleave iron from extracellular ferritin’s protein shell, allowing “free iron” to participate in FR generation.13–16 The alveolar macrophage (AM) attempts to protect against these potential iron-catalyzed oxidative insults by scavenging iron and sequestering it as chemically inert hemosiderin and ferritin.17 Hemosiderin-laden macrophages (HLMs) are often noted anecdotally in routine BAL and TBB specimens from LT recipients and are usually considered (although with little, if any evidence) to be secondary to bleeding after previous biopsy procedures.18 However, we hypothesize that the presence of numerous HLMs in bronchoalveolar lavage samples following LT is a surrogate marker of a more generalized iron load. Increased microvascular permeability secondary to airway and/or lung parenchymal inflammation may be another potential source of iron within the
Reid et al.
841
airway in LT recipients. Although microvascular leakage is not well described in LT, the inflammatory processes involved in liberating ferritin and iron from cellular sources (i.e., NO and tumor necrosis factor-␣ [TNF-␣] production) have also been shown to increase pulmonary permeability, and similar mechanisms could be operating within the allograft.19,20 BALF albumin is elevated in some subjects with non–lung-transplant-related pulmonary inflammation,21,22 and in this study we have determined its concentration as a marker of lung vascular leakage following LT. To test our hypothesis that the lung allograft is subject to significant iron and NO-generated oxidative stress, we measured the concentrations of ferritin and the stable NO derivative nitrite (NO2⫺) in bronchoalveolar lavage fluid (BALF) of ST and BOS subjects compared with healthy, non-smoking normal controls.23 We formally quantified HLMs from LT recipients and assessed their relationship to BALF ferritin. We also determined BALF albumin concentrations and differential cell counts to determine whether increased pulmonary vascular permeability and polymorphonuclear cells (PMNs) might be relevant to the disease processes that contribute to the oxidative load.
MATERIALS AND METHODS Subjects and Equipment BALF samples from 21 subjects who had undergone LT at the Alfred Hospital between March 1991 and October 1994 were studied (Tables I and II). Among the study population, presence of BOS was defined as a “fixed” 20% or greater fall in forced expiratory volume in 1 second (FEV1) from the best achieved post-transplant in the absence of any other identifiable cause. Stable transplant recipients were defined as those who were clinically well, had well-preserved lung function (i.e., within 20% of their best posttransplant FEV1) and no more than A1 grading for acute lung rejection on standard histologic grading of trans-bronchial lung biopsies.24 Bronchoscopy was performed as part of routine surveillance or for investigation of deteriorating lung function in the BOS group. No bronchoscopic procedure included in this study was performed within 4 weeks of the last, and the median time from the previous procedure was 87 days (range 28 to 912 days). The transplant population studied consisted of 14 STs (median age 34 years, range 20 to 46 years; 8 women; median time from transplant 178 days, range 65 to 1,128 days) and 7 subjects with BOS
842
Reid et al.
The Journal of Heart and Lung Transplantation August 2001
TABLE I Clinical characteristics of stable lung transplant recipients Subject
Age (years)
Gender
Original Diagnosis
Operation
Time From Tx (Days)
Time From TBB (Days)
FEV1 liters/second
FEV1% Max
1 2 3 4 5 6 7 8 9 10 11 12 13 14
46 41 41 32 33 22 34 33 46 40 20 38 22 48
M F F F F F F M F F M M M F
Sarcoid PPH E CF E CF E E B B E CF CF COPD
BSLT HLT HLT BSLT HLT BSLT HLT HLT BSLT BSLT HLT BSLT BSLT (R)SLT
84 125 534 65 427 101 178 1128 108 178 410 542 84 375
29 73 178 44 101 37 93 359 63 87 199 186 28 84
3.4 2.5 3.4 1.8 3.6 3.4 3.5 3.4 1.9 1.9 4.3 3.2 3.1 2.0
100 100 100 95 100 100 100 100 100 100 99 91 100 92
CF, cystic fibrosis; PPH, primary pulmonary hypertension; E, Eisenmenger’s syndrome; ␣ 1-AT, ␣ 1-anti-trypsin deficiency; B, bronchiectasis; COPD, chronic obstructive pulmonary disease; BSLT, bilateral sequential lung transplant; HLT, heart–lung transplant; SLT, single lung transplant; L/R, left/right; Tx, transplant; TBB, trans-bronchial biopsy; FEV1% max, forced expiratory volume at 1 sec as a percentage of best achieved result post-transplant.
(median age 25 years, range 19 to 50 years; 4 women; median time from transplant 1,080 days, range 270 to 1,440 days). All BALF aspirates were routinely evaluated for evidence of viral, bacterial or fungal infection by conventional microscopic and culture techniques and, if significant infection was thought to be present, the sample was excluded from study. Twenty-one healthy, non-smoking volunteers (median age 23 years, range 19 to 38 years) were recruited to provide control data for BALF indices. All control subjects had normal lung function and were not taking any regular medications. The study was approved by our institutional ethics committee and written informed consent was obtained from each subject.
Bronchoscopic Procedure Flexible fiber-optic bronchoscopy was performed using topical anesthetic (lignocaine 4% above the cords and 2% below in 2-ml aliquots, total dose ⱕ 24 mg) and intravenous midazolam as a sedative. The bronchoscope was passed trans-nasally and wedged in a sub-segmental bronchus of either the right middle lobe or the lingula segment of the left upper lobe. Three 60-ml aliquots of sterile, phosphatebuffered saline solution, pre-warmed to 37°C, were injected and aspirated immediately with moderate suction (⫺50 to ⫺100 mm Hg) after each individual administration. The BALF aspirated was immediately placed into siliconized glass vessels on ice.
TABLE II Clinical characteristics of lung transplant recipients with bronchiolitis obliterans syndrome Subject
Age (Years)
Gender
Original Diagnosis
Operation
Time From Tx (Days)
Time From TBB (Days)
FEV1 liters sec
FEV1% Max
15 16 17 18 19 20 21
25 24 19 20 42 30 50
M M M M F F F
CF CF CF CF ␣ 1-AT PPH COPD
BSLT BSLT HLT HLT (L)SLT BSLT (R)SLT
270 540 1080 1080 1440 395 1100
72 30 226 218 912 54 464
2.52 1.3 1.9 2.6 0.6 0.5 0.7
75 55 64 52 28 21 49
CF, cystic fibrosis; PPH, primary pulmonary hypertension; E, Eisenmenger’s syndrome; ␣ 1-AT, ␣ 1-anti-trypsin deficiency; B, bronchiectasis; COPD, chronic obstructive pulmonary disease; BSLT, bilateral sequential lung transplant; HLT, heart–lung transplant; SLT, single lung transplant; L/R, left/right; Tx, transplant; TBB, trans-bronchial biopsy; FEV1% max, forced expiratory volume at 1 sec as a percentage of best achieved result post-transplant.
The Journal of Heart and Lung Transplantation Volume 20, Number 8
Reid et al.
843
Cell Staining and Counting The volume of BAL return was recorded and the fluid filtered through a 200-m-pore-size stainlesssteel filter. Total cell count was performed on unprocessed BALF using a hemocytometer. BALF cytospins were prepared by centrifuging at 850 rpm for 10 minutes using a Shandon II cytocentrifuge and stored at ⫺80°C, so that subsequent counting could be done in standardized runs. BALF was sub-divided into 1.8-ml aliquots and stored at ⫺80°C. No part of the aspirate was discarded. Stored frozen cytospots were subsequently taken from the freezer and thawed at room temperature for 30 minutes, then aired with methanol for 10 minutes and immediately washed with distilled water. Cytospots were stained with Perl’s reagent for 20 minutes, washed and counterstained with nuclear-fast red for 10 minutes. Slides were then washed, dried and coverslipped. A hemosiderin score (HS) was calculated as described by Kahn.25 Briefly, 200 macrophages were examined on each slide and each cell was ranked for hemosiderin content using the following scale: 0 ⫽ no color; 1 ⫽ faint blue in one portion of cytoplasm; 2 ⫽ deep blue in a minor portion of the cell; 3 ⫽ deep blue in most areas of the cytoplasm; and 4 ⫽ deep blue throughout the cell. The total value for all cells was calculated and divided by 2 to obtain a score for an average of 100 cells. In addition, the simple percentage of cells staining positively was also recorded. Cell counting and scoring was performed by one experienced observer blinded to subject category.
BALF Ferritin Ferritin concentration in cell-free, unconcentrated BALF was determined by fluorescence immunoassay using an AXSYM analyzer and expressed as micrograms per liter.
BALF Albumin Albumin concentration was assessed using an autoanalyzer (Cobas Fara II, Roche Diagnostics, Basel, Switzerland) and expressed as milligrams per liter.
FIGURE 1 Hemosiderin score for stable lung
transplant recipients (ST), subjects with BOS and normal controls.
2.6 mol/liter NaOH was added to maximize fluorescence. Fluorescence of the final reaction product, 1(H)-naphotriazole (NAT), was determined using a spectrometer (Aminco-Bowman, Rochester, NY) and results expressed as the (nano)molar concentration of nitrite.
Statistical Methods Results are expressed as medians and ranges. The Kruskal–Wallis test was used to compare the three subject groups and the Mann–Whitney test was used to evaluate differences between specific groups. Spearman’s rank correlation was used to examine relationships between variables. Analysis was performed using the MINITAB for WINDOWS statistics program. A two-tailed p ⬍ 0.05 was considered significant.
BALF Nitrite (DAN Assay)
RESULTS Hemosiderin Score (HS) and BAL Cell Differentials
BALF NO2⫺ concentrations were determined by a fluorometric assay using 2,3-diaminonaphthalene (DAN; Molecular Probes, Eugene, OR). One hundred microliters of DAN solution was added to each 400 l of BAL sample. After thorough mixing the sample was incubated in the dark for 10 minutes and
BAL cells from 19 of the 21 transplant subjects and 3 of the 11 controls examined stained positively for hemosiderin within macrophages (Figure 1 and Table III). HS was significantly higher in both ST (median 10.5, range 0 to 187) and BOS subjects (median 6.5, range 0 to 27) compared with normal
844
Reid et al.
The Journal of Heart and Lung Transplantation August 2001
TABLE III Bronchoalveolar lavage findings in lung transplant recipients and normal subjects
ST BOS Normal controls
HS
BALF Neutrophils %
BALF Ferritin (g/liter)
BALF Albumin (mg/liter)
BALF Nitrite (nmol/liter)
10.5a (0–187) 6.5b (0–27) 0 (0–2)
4.0c (0.6–17.7) 46.4d (15.8–87.4) 1.2 (1.0–3.3)
48.8e (10–950) 126f (1–205) 2 (0–16)
33.8 (7–328) 57.9g (34.1–309) 30.2 (17.8–63.4)
440h (256–1194) 537i (443–1044) 337 (256–531)
BALF, bronchoalveolar lavage fluid; BOS, bronchiolitis obliterans syndrome; HS, hemosiderin score; ST, stable transplants. Hemosiderin score aST and bBOS vs normal, p ⫽ 0.001 and p ⫽ 0.01, respectively. BALF neutrophils %: cST and dBOS vs normal, p ⫽ 0.008 and p ⬍ 0.001, respectively (BOS vs ST. p ⬍ 0.001). BALF ferritin: eST and fBOS vs normal, p ⬍ 0.001 and p ⫽ 0.03, respectively. BALF albumin: g BOS vs normal and BOS vs ST, p ⫽ 0.02 and p ⫽ 0.05, respectively. BALF nitrite: hST and iBOS v normal, p ⫽ 0.09 and p ⫽ 0.002, respectively.
controls (median 0, range 0 to 2; p ⫽ 0.001 and p ⫽ 0.01, respectively). The percentage of cells positive for hemosiderin was also significantly higher in ST (median 7.5%, range 0% to 74.5%) and BOS subjects (median 5.0%, range 0% to 10.0%) than in normal controls (median 0%, range 0% to 1.5%; p ⬍ 0.001). There was no significant difference in HS or percentage of positive cells between LT recipient groups, that is, with or without BOS (p ⫽ 0.3). BAL neutrophils were significantly elevated in both ST (median 4.0%, range 0.6% to 17.0%) and BOS subjects (median 46.4%, range 15.8% to 87.4%), compared with normal controls (median 1.2%, range 1.0% to 3.3%; p ⫽ 0.008 and p ⬍ 0.001, respectively), and there was a significant difference between the two LT recipient groups (p ⬍ 0.001).
BALF Nitrite BALF NO2⫺ concentrations were significantly higher in LT recipients overall (median 475 nmol/ liter, range 256 to 1,194 nmol/liter) compared with normal controls (median 337 nmol/liter, range 256 to 531 nmol/liter; p ⬍ 0.05), but there was no significant difference between LTR groups (Figure 4 and Table III).
Correlations There was a significant correlation between BALF ferritin concentration and HS (Figure 5) in LTR
BALF Ferritin BALF ferritin concentrations were significantly higher in all LT recipients overall (median 49.0 g/liter, range 1 to 950 g/liter) compared with normal controls (median 2 g/liter, range 0 to 16 g/liter, p ⬍ 0.01), but there was no significant difference between ST (median 48.8 g/liter, range 10 to 950 g/liter) and BOS subjects (median 126 g/liter, range 1 to 205 g/liter, p ⫽ 0.9) (Figure 2 and Table III).
BALF Albumin BALF albumin levels were significantly higher in BOS subjects (median 57.9 g/liter, range 34.1 to 309 g/liter) compared with normal controls (median 30.2 g/liter, range 17.8 to 63.4 g/liter; p ⫽ 0.02) and there was a strong trend toward higher levels compared with ST subjects (median 33.8 g/liter, range 7.0 to 328 g/liter; p ⫽ 0.05) (Figure 3 and Table III). However, there was no difference in BALF albumin between ST subjects and normal controls (p ⫽ 0.8).
FIGURE 2 BALF ferritin concentrations in stable lung transplant recipients (ST), subjects with BOS and normal controls.
The Journal of Heart and Lung Transplantation Volume 20, Number 8
Reid et al.
845
FIGURE 3 BALF albumin concentrations in stable
lung transplant recipients (ST), subjects with BOS and normal controls.
overall (r(s) ⫽ 0.7, p ⬍ .001). This relationship was much stronger in ST (r(s) ⫽ 0.7, p ⬍ 0.005) than BOS subjects (r(s) ⫽ 0.4, p ⫽ 0.3), although this probably reflects the small sample size and lack of statistical power in the latter group. Similarly, the relationship between BALF ferritin concentration and %hemosiderin-positive AM was stronger in ST (r(s) ⫽ 0.8, p ⬍ 0.005) than BOS subjects (r(s) ⫽ 0.6, p ⫽ 0.2), although as a group a significant relationship remained in LT recipients overall (r(s) ⫽ 0.7, p ⬍ 0.001). BALF ferritin and albumin concentrations were significantly related in BOS subjects (r(s) ⫽ 0.8, p ⫽ 0.05) despite the small numbers, but there was no such relationship in ST subjects. Similarly, there was a weak relationship between BALF ferritin and nitrite concentrations in BOS (r(s) ⫽ 0.6, p ⫽ 0.1), but not in ST. Furthermore, there was a strong correlation between BALF NO2⫺ concentration and BAL %neutrophils in BOS subjects (r(s) ⫽ 0.9, p ⬍ 0.001), but not in ST subjects.
Clinical Outcomes There was no relationship between HS and %hemosiderin-positive cells or BALF ferritin concentration and time from transplant or previous biopsy procedure. All 13 ST cases remained well during the initial 6 months following the study procedures, but 5 sub-
FIGURE 4 BALF nitrite concentrations in stable lung transplant recipients (ST), subjects with BOS and normal controls.
sequently developed BOS during long-term follow-up (median time from bronchoscopy 44 months, range 34 to 70 months), and 3 of these have died. Four of the 7 patients with BOS died during followup. There was no obvious relationship between HS and/or BALF solute indices at the time of this reported sampling and subsequent outcomes in either the ST or BOS groups.
DISCUSSION The current study demonstrates a significant increase in BALF ferritin concentration and HLM numbers in the lower respiratory tract of lung transplant recipients (LTRs). We believe these findings are very unlikely to just be the sequelae of previous biopsy procedures. In addition, we found elevated levels of nitrite in BALF of LTRs, suggesting the allograft may be subject to significant iron and NO-derived free-radical damage. Furthermore, we confirmed previous observations that BAL %neutrophils are increased following LT, even in apparently stable subjects, which suggests that neutrophil-mediated airway inflammation may persist despite apparent well-being and intense immunosuppression and could be relevant in the pathogenesis of chronic rejection. Subjects with BOS had significantly higher levels of BALF albumin com-
846
Reid et al.
The Journal of Heart and Lung Transplantation August 2001
FIGURE 5 Relationship between BALF ferritin concentration and hemosiderin score in all lung transplant recipients (i.e., ST and BOS) as a group.
pared with normal controls and there was a strong trend for higher levels compared with ST. In addition, there was a significant relationship between BALF albumin and ferritin levels in BOS, which may indicate a microvascular leak syndrome that could contribute to the presence of ferritin and HLM within the airway. Our findings add to the literature concerning the potential role of iron overload and oxidative stress following LT. A recent study by Baz et al6 demonstrated elevated levels of total iron in the putative epithelial lining fluid (ELF) of lung allografts as well as an increase in tissue iron deposition in a small number of stable LTRs. We confirmed these observations in a larger population of patients and demonstrated similar findings in subjects with established BOS. An excess of chemically available iron within the lower respiratory tract may promote the catalysis of toxic hydroxyl radicals, which could contribute to the pathogenesis of cell injury and airway inflammation following LT.6,13,26 Also, the fibrogenic influence of iron is well recognized in overload syndromes such as hemochromatosis and idiopathic pulmonary hemosiderosis, and an increase in lung iron may be part of a final common pathway that ultimately leads to growth factor gen-
eration, promotion of collagen gene transcription and airway fibrosis and remodeling.26 Under normal conditions, iron within the lower respiratory tract is present in extremely low concentrations, bound to ferritin and other iron-binding proteins that prevent hydroxyl radical formation and limit the availability to iron-requiring microorganisms.26 However, the lung allograft may be subject to a number of inflammatory stimuli that can disrupt iron homeostasis and increase its chemical availability. Intracellular ferritin mRNA transcription and production can be up-regulated by TNF-␣, neutrophil-derived hydrogen peroxide and NO, all of which have been implicated in the processes of reperfusion injury, acute rejection and frequent infections following LT.12–15,27,28 A generalized increase in ferritin synthesis and storage could function as a reservoir of potentially reactive iron if intracellular stores are released secondary to cell injury or death. Extracellular ferritin will release its core of iron atoms in the presence of PMN and NO-derived radicals and our finding of increased PMN numbers and nitrite concentrations, in the setting of elevated BALF ferritin, suggests the allograft is likely to be subject to significant iron toxicity.16,29,30
The Journal of Heart and Lung Transplantation Volume 20, Number 8
To protect against these oxidative stresses the lung has evolved several defense mechanisms, which prominently involve the AM. These produce numerous anti-oxidant compounds, such as metallothionein and glutathione, which are detectable within ELF following LT and sequester iron intracellularly as chemically inert hemosiderin and ferritin.17,31,32 HLMs are seen following inhalation of iron-containing dusts, after alveolar hemorrhage and in cigarette smokers, but they have also been described in pulmonary fibrosis, suggesting they may be involved in the broader processes of lung inflammation and remodeling.17,18,33 Paradoxically, increased uptake and storage of iron within AMs may, under certain circumstances, enhance iron toxicity, particularly if intracellular ferritin synthesis lags behind the rate of iron uptake.34 This may be of particular relevance in LT where the ability of the AMs to leave the allograft is impaired because of diminished mucociliary clearance and lymphatic drainage. Thus, iron sequestration in lung allografts by AMs may be a “double-edged sword” and the relationship between BALF ferritin and HLMs in our study could be especially harmful if it reflects continuous cycling of ferritin-bound iron that may well have been released by injured bronchial epithelial cells and senescent oxidant-damaged AMs. Indeed, we have current data, as yet unpublished, from FACS analysis, which suggests that AMs are functionally impaired in LTR, and iron toxicity may be a contributing factor.35 The unknown lifespan of an AMs in lung allografts raises the potential criticism that the presence of HLMs within BALF in our study is simply representative of accumulated iron secondary to hemorrhage from previous biopsy procedures. We believe this to be unlikely for a number of reasons. First, BALF was obtained deliberately from either the right middle lobe or lingula segment of the left upper lobe specifically to avoid the sites ordinarily used for biopsy purposes at our center. In addition, despite the uncertainty concerning the time required for HLMs to leave the airway their lifespan probably remains limited and, under normal conditions, could be expected to disappear within 2 weeks of hemorrhage.18,36 In our study no TBB was undertaken within 28 days of that most recently performed procedure and there was absolutely no relationship between time from transplant or previous biopsy and BALF ferritin concentration or HS, which suggests that bleeding from biopsy procedures is not relevant. It is also
Reid et al.
847
highly improbable that the elevated levels of BALF ferritin in LTRs (median 49 ⫻ 10⫺6 g/liter) in our study are related to previous alveolar hemorrhage. This is because red blood cells (RBCs) contain extremely low concentrations of ferritin (⬍10⫺19 g/liter) and, according to our calculations, it would require a blood loss of at least 10 liters to account for the ferritin detected if this were the only source; that is, 10 (liters) ⫻ 5 ⫻ 10⫺12 (RBCs per liter of blood) ⫻ 10⫺19 (RBC ferritin content, in micrograms per liter) ⫽ 50 ⫻ 10⫺6 g/liter. Furthermore, although vascular leakage per se may be responsible for some of the ferritin detected, plasma exudation either as a consequence of airway inflammation or the lavage procedure itself is unlikely to be a significant contributor. A crude estimate of the “true” ELF concentration of ferritin can be calculated if one assumes a lavage dilution factor of at least 100, which equates to a median ELF ferritin concentration of 49 ⫻ 10⫺4 g/liter, which clearly greatly exceeds normal plasma levels (15 to 300 ⫻ 10⫺6 g/liter). In this study we were unable to demonstrate a significant increase in BALF albumin in LTRs as a whole compared with controls, but the elevated BALF albumin levels in BOS subjects probably does represent an increase in microvascular leakage. The latter could be occurring as a consequence of NO acting upon the vascular endothelium, although the absence of a direct relationship between BALF nitrite and albumin concentrations suggests other factors are also operative. Furthermore, although the trend toward a relationship between BALF ferritin and albumin in LTRs overall and a significant relationship in BOS subjects suggests vascular leakage may be responsible for some of the ferritin seen, it is perhaps not too surprising in our study that an isolated BALF albumin correlated poorly with HS, because of the difference in temporal history of their relative accumulations in the allograft. In conclusion, evidence is accumulating to support the hypothesis that the lung allograft is subject to significant oxidative stress mediated by neutrophil production of ROS and probably exacerbated by NOderived radicals, and especially the presence of an excess of tissue iron. The allograft may have an impaired ability to combat these insults because of antioxidant depletion and cell damage continuing unabated, with the eventual development of chronic rejection, airway remodeling and ultimately organ failure.31,32,37–39 The presence of ferritin and HLMs in BALF does not merely represent hemorrhage after
848
Reid et al.
recent lung biopsies, but rather reflects a generalized iron overload that may be of etiologic significance in the evolution of chronic rejection. Therapeutic interventions might involve dietary supplementation with anti-oxidants or iron chelation to reduce the oxidative burden and animal studies of acute rejection in cardiac and renal transplantation have demonstrated the benefits of such potential treatments.41,42 Prospective studies now need to be undertaken to address these issues and examine the relationship between allograft iron deposition, oxidative stress and anti-oxidant depletion and to relate these interactions to the development of chronic rejection. REFERENCES 1. Valentine VG, Robbins RC, Berry GJ, Patel HR, Reichenspurner H, Reitz BA, Theodore J. Actuarial survival of heart–lung and bilateral sequential lung transplant recipients with obliterative bronchiolitis. J Heart Lung Transplant 1996;15:371– 83. 2. Yousem SA, Paradis IL, Dauber, JH, et al. Efficacy of transbronchial lung biopsy in the diagnosis of bronchiolitis obliterans in heart–lung transplant recipients. Transplantation 1989;47:893–5. 3. Yousem SA, Suncan SR, Ohori P, Sonmez-Alpan E. Architectural remodelling of lung allografts in acute and chronic rejection. Arch Pathol Lab Med 1992;116:1175– 80. 4. Snell GI, Esmore DS, Williams TJ. Cytolytic therapy for the bronchiolitis obliterans syndrome complicating lung transplantation. Chest 1996;109:874 – 8. 5. Riise, GC, Williams A, Kjellstrom C, Schersten H, Andersson BA, Kelly FJ. Bronchiolitis obliterans syndrome in lung transplant recipients is associated with increased neutrophil activity and decreased anti-oxidant status in the lung. Eur Resp J 1998;12:82– 8. 6. Britton RS, Tavill AS, Bacon BR. Mechanisms of iron toxicity. In: Brock JH, Halliday JW, Pippard MJ, Powell LW, eds. Iron metabolism in health and disease. London: Saunders; 1994:311–51. 7. Freeman B. Free radical chemistry of nitric oxide. Looking at the dark side. Chest 1994;105(Suppl):79S– 84S. 8. Saleh D, Ernst P, Lim S, Barnes PJ, Giaid A. Increased formation of the potent oxidant peroxynitrite in the airways of asthmatic patients is associated with induction of nitric oxide synthase: effect of inhaled glucocorticoid. FASEB J 1998;12:929 –37. 9. Baz MA, Ghoi AJ, Roggli VL, Tapson VF, Piantadosi C. Iron accumulation in lung allografts after transplantation. Chest 1997;112:435–9. 10. Giaid A, Corris PA, Chikhani N, et al. Expression of nitric oxide synthase in lung transplant recipients with bronchiolitis obliterans. Eur Respir J 1995;8(Suppl):550S. 11. Gabbay E, Walters EH, Orsida B, et al. In stable lung transplant recipients, exhaled nitric oxide levels positively correlate with airway neutrophilia and bronchial epithelial iNOS. Am Rev Respir Crit Care Med 1999;160:2093–9. 12. Devlin J, Palmer RM, Gonde CE, et al. Nitric oxide generation: a predictive parameter of acute allograft rejection. Transplantation 1996;61:745–9.
The Journal of Heart and Lung Transplantation August 2001 13. Gutteridge JMC, Mumby S, Quinlan GJ, Chung KF, Evans TF. Pro-oxidant iron is present in human pulmonary epithelial lining fluid; implications for oxidative stress in the lung. Biochem Biophys Res Commun 1996;220:1024 –7. 14. Miller CC, Miller SC, Torti SV, Tsuji Y, Torti FM. Ironindependent induction of ferritin H-chain by tumor necrosis factor. Proc Natl Acad Sci USA 1991;88:4946 –50. 15. Kwak EL, Larochelle DA, Beaumont C, et al. Role for NF- in the regulation of ferritin H by tumour necrosis factor. J Biol Chem 1995;270:15285–93. 16. Reif DW. Ferritin as a source of iron for oxidative damage. Free Rad Biol Med 1992;12:417–27. 17. Corhay JL, Weber G, Bury T, et al. Iron content in human alveolar macrophages. Eur Respir J 1992;5:804 –9. 18. Grebski E, Hess T, Hold G, Speich R, Russi E. Diagnostic value of haemosiderin laden macrophages in bronchoalveolar lavage. Chest 1992;102:1794 –9. 19. Bernareggi M, Mitchell JA, Barnes PJ, Belvisi MG. Dual action of nitric oxide on airway plasma leakage. Am J Crit Care Med 1997;155:869 –74. 20. Stephens KE, Ishizaka A, Larrick JW, Raffin TA. Tumour necrosis factor causes increased pulmonary permeability and edema. Am Rev Respir Dis 1987;137:1364 –70. 21. Jones KP, Edwards JH, Reynolds SP, Peters TJ, Davies BH. A comparison of albumin and urea as reference markers in bronchoalveolar lavage fluid from patients with interstitial lung disease. Eur Resp J 1990;3:152– 6. 22. Fick RB, Metzger JW, Hal RB, et al. Increased bronchovascular permeability after allergen challenge in sensitive asthmatics. J Appl Physiol 1987;63:1147–55. 23. Tayeh MA, Marletta MA. Macrophage oxidation of Larginine to nitric oxide, nitrite, and nitrate. J Biol Chem 1989;33:19654 – 8. 24. Yousem SA, Novick RJ, Breen TJ, et al. A working formulation for the standardisation of nomenclature in the diagnosis of heart and lung rejection: lung rejection study group. J Heart Transplant 1990;9:593– 601. 25. Kahn FW, Jones JM, England DM. Diagnosis of pulmonary haemorrhage in the immunocompromised host. Am Rev Respir Dis 1987;136:155– 60. 26. Mateos F, Brock JH, Perez-Arellano JL. Iron metabolism in the lower respiratory tract. Thorax 1998;53:594 – 600. 27. Silkoff PE, Caramori M, Tremblay L, et al. Exhaled nitric oxide in human lung transplantation; a non-invasive marker of acute rejection. Am J Respir Crit Care Med 1998;157:1822– 8. 28. Gabbay E, Walters EH, Orsida B, et al. Post lung transplant bronchiolitis obliterans syndrome (BOS) is characterized by increased exhaled nitric oxide levels and epithelial iNOS. Am J Respir Crit Care Med 2000;162:2182–7. 29. Pantopoulos K, Weiss G, Hentze MW. Nitric oxide and oxidative stress control mammalian iron metabolism by different pathways. Mol Cell Biol 1996;16:3781– 8. 30. Reif DW, Simmons RD. Nitric oxide mediates iron release from ferritin. Arch Biochem Biophys 1990;283:537– 41 31. Courtade M, Carrera G, Paternain JL, et al. Metallothioein expression in human lung and its varying levels after lung transplantation Chest 1998;113:371– 8. 32. Baz MA, Tapson VF, Roggli VL, Van Trigt P, Piantadosi CA. Glutathione depletion in epithelial lining fluid of lung allograft patients. Am J Respir Crit Care Med 1996;153: 742– 6. 33. Rich EA. Iron metabolism by smokers’ alveolar macro-
The Journal of Heart and Lung Transplantation Volume 20, Number 8
34.
35.
36.
37.
38.
phages: protective or problematic in the lung? J Lab Clin Med 1988;111:598 –9. Wesselius LJ, Williams WL, Bailey K, et al. Iron uptake promotes hyperoxic injury to alveolar macrophages. Am Rev Respir Crit Care Med 1999;159:100 – 6. Ward C, Whitford H, Snell GI, et al. Bronchoalveolar lavage macrophage and lymphocyte phenotypes in lung transplant recipients. J Heart Lung Transplant. In press. Sherman MJ, Winnie G, Thomassen MJ, Abdul-Karim FW, Boat TF. Time course of haemosiderin formation and clearance by human pulmonary macrophages. Chest 1984;86:409 –11. Shivakumar BR, Kolluri SV, Ravindranath V. Glutathione homeostasis in brain during reperfusion following bilateral carotid artery occlusion in the rat. Mol Cell Biochem 1992; 111:125–9. Riise GC, Williams A, Kjellstrom C, Schersten H, Andersson BA, Kelly FJ. Bronchiolitis obliterans syndrome in lung
Reid et al.
39.
40.
41.
42.
849
transplant recipients is associated with increased neutrophil activity and decreased anti-oxidant status in the lung. Eur Respir J 1998;12:82– 8. Kobayashi H, Nonami T, Kurokawa S, et al. Changes in the glutathione redox system during ischaemia and reperfusion in rat liver. Scand J Gastroenterol 1992;27:711– 6. Martelius T, Scholz M, Krogerus L, et al. Antiviral and immunomodulatory effects of desferrioxamine in cytomegalovirus-infected rat liver allografts with rejection. Transplantation 1999;68:1753– 61. Slakey DP, Roza AM, Pieper GM, et al. Delayed cardiac allograft due to combined cyclosporine and antioxidant therapy. Transplantation 1993;56:1305–9. Wang C, Salahudeen AK. Cyclosporine nephrotoxcity: attenuation by an antioxidant inhibitor of lipid peroxidation in vitro and in vivo. Transplantation 1994;58:940 – 6. the previous procedure was 87 days (range 28 to 912 days).