Surfactant protein-A (SP-A) leaks into the circulation of patients with acute respiratory distress syndrome (ARDS) or acute cardiogenic pulmonary edema (APE) in a manner inversely related to lung function. Since surfactant protein-B (SP-B) is synthesized as a precursor considerably smaller than alveolar SP-A, we investigated whether it enters the circulation more readily. Reactivities consistent with SP-B proprotein ( ∼ 42 to ∼ 45 kD) and the ∼ 25 kD processing intermediate were detected in plasma. Plasma immunoreactive SP-B levels were significantly higher in ARDS (8,007 ± 1,654 ng/ml [mean ± SEM], n = 22) and APE (3,646 ± 635 ng/ml, n = 10) patients compared with normal subjects (1,685 ± 58 ng/ml, n = 33) and ventilated patients with no cardiorespiratory disease (1,829 ± 184 ng/ml, n = 7). All groups had plasma SP-B/SP-A ratios ∼ 6- to ∼ 8-fold higher than in normal lavage or ARDS tracheal aspirate fluid, consistent with protein sieving. During admission, both plasma SP-B and the SP-B/SP-A ratio were inversely related to blood oxygenation (PaO2 /Fi O2 ) (p < 0.0001 and p < 0.025, n = 260 from 39 patients; Spearman) and static respiratory system compliance ( Δ V/ Δ P) (p < 0.0001 and p < 0.01, n = 168 from 25 patients). We describe in detail three patients and conclude that immunoreactive SP-B enters more readily than SP-A, is cleared acutely, and provides a better indicator of lung trauma.
The hallmark of the acute respiratory distress syndrome (ARDS) is an increase in the alveolocapillary permeability arising from insults via the airways or the blood. In most cases, a common pathway involving the recruitment and activation of leukocytes is thought to result in the diffuse inflammatory lung injury. Following the loss of the macromolecular barrier, protein-rich edema fluid floods the alveoli and impairs surfactant function, with ensuing atelectasis and a concomitant deterioration in gas exchange and lung compliance. Although ventilation aimed at maintaining arterial oxygen content is standard treatment for critically ill patients with respiratory failure, these regimens may aggravate the existing high-permeability edema characteristic of the injured lung (1). The issue of which pressure, peak dynamic, peak static, mean, or other, best maintains blood oxygenation and cardiac output, while minimizing lung trauma, is highly contentious. Hence, detection and quantification of early damage to the alveolocapillary membrane has important implications for treatment modalities.
Although surfactant proteins are normally only found in appreciable amounts in the lung, leakage of surfactant protein-A (SP-A) (2-4), surfactant protein-B (SP-B) (3), and surfactant protein-D (SP-D) (5) into the circulation has been reported in a number of respiratory disorders. The route by which the proteins enter the circulation is unknown; however, there is strong evidence that a bidirectional plasma protein flux occurs in lungs, the magnitude of which depends on disease severity (6). Recently, we reported elevated concentrations of SP-A in the serum of patients with ARDS and in those with acute cardiogenic pulmonary edema (APE) (7), where stress failure from excessively high pulmonary capillary pressures may have damaged the alveolocapillary membrane. Moreover, serum SP-A was inversely related to blood oxygenation and static respiratory system compliance.
In the airspaces, SP-A predominantly forms high-molecular-weight oligomers (∼ 650 kD) with Stokes radii of ∼ 35 nm (8). Although mature SP-B, which associates as a low-Mr (∼ 18 kD), thiol-dependent homo-dimer (9), is normally intimately associated with complexes of surfactant phospholipids, possibly too large to readily breach the alveolocapillary barrier (10), labeling studies in isolated type II cells suggest that at least some of the protein is secreted into the alveolus as hydrophilic, monomeric proprotein and processing intermediate with Mr of ∼ 45 kD and ∼ 25 kD, respectively (11). Since these immunoreactive SP-B forms are considerably smaller than native SP-A, they possibly breach the alveolocapillary barrier more readily and so may provide a better marker of lung injury.
We have now compared the plasma concentration of immunoreactive SP-A and -B in our original cohort (7) and in some newly recruited patients. In addition, we describe in detail three ventilated patients: a patient without respiratory disease ventilated for 9 d, and ARDS patient who recovered from severe lung injury, and an ARDS patient who died from respiratory failure.
The study was approved by the Ethics Review Committee for Clinical Investigation of the Flinders Medical Centre (Permit No. 26/93). Informed consent for blood sampling was obtained directly from the subjects or from the closest relative in the case of most of the ventilated patients.
Blood sampling. Blood was collected from: (1) 33 subjects (21 males, 12 females; median age: 28 yr [range: 17 to 50 yr]) with no history of cardiopulmonary disease (normal subjects). (2) Seven patients (four males, 3 females; median age: 20 yr [range: 15 to 39 yr]) with no history or current evidence of cardiorespiratory disease (OD) (7). Six of the patients required mechanical ventilation because of drug overdose and were intubated for at least 6 h before blood was sampled. Blood was sampled daily from the seventh patient who was mechanically ventilated while awaiting emergency orthotopic liver transplantation as a consequence of hepatic encephalopathy due to Wilson's disease. (3) Ten patients (five males, 5 females; median age: 74 yr [range: 63 to 83 yr]), all of whom had dyspnea of sudden onset and a clinical course consistent with a diagnosis of APE (7). Blood was sampled early in the course of the disease and again ∼ 24 h later during its resolution. (4) Twenty-two consecutive consenting patients (13 males, eight females; median age: 64 yr [range: 21 to 80 yr]) requiring mechanical ventilation for ARDS. All patients had a lung injury score ⩾ 2.5 at the time of entry into the study when ARDS was diagnosed (7). Blood was drawn ∼ 7:00 a.m. on each day in the intensive care unit (ICU). A matching mixed venous blood was collected via an indwelling pulmonary artery catheter from 11 of the patients.
Plasma preparation. Blood was drawn from an antecubital vein from normal subjects and from an indwelling arterial catheter in the case of ventilated patients and was immediately centrifuged in lithium heparin tubes at 5,000 rpm for 5 min at room temperature (Megafuge; Heraeus-Christ, Osterode, Germany) and the plasma stored at −20° C for batch analysis.
Clinical measurements. Whenever possible, pulmonary capillary pressure (PCP) and static respiratory system compliance (ΔV/ΔP) were measured each time blood was sampled (7). The clinical variables of the original patient cohort have been summarized elsewhere (7). The clinical variables of the additional patients, both on Day 1 and immediately before their discharge from the ICU or their death, are summarized in Table 1.
Patient | Age/Sex | Days in Study (n) | Lung Injury Score/Insult | PEEP (cm H2O) | Static ΔV/ΔP (ml/cm H2O) | Fi O2 | PaO2 /Fi O2 | |||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
OD | 20/M | 0/9 | 0.4/Wilson's | 0.0:3.0 | ND:63.2 | 0.4:0.25 | 420:476 | |||||||
ARDS | ||||||||||||||
33 | 35/M | 0/25 | 2.75/pneumonia | 7.5:10.0 | 51.4:ND | 1.0:0.5 | 69:213 | |||||||
34 | 26/M | 0/11 | 3.00/pneumonia | 12.0:5.0 | 40.8:25.0 | 0.9:0.4 | 108:275 | |||||||
35 | 66/F | 0/8 | 3.00/sepsis | 10.0:0.0 | 32.0:ND | 0.7:0.4 | 169:217 | |||||||
36 | 64/M | 3/21 | 2.75/aspiration | 8.0:12.5 | 54.9:49.6 | 0.8:1.0 | 91:68 | |||||||
37 | 74/M | 4/2 | 2.75/aspiration | 10:10 | 44.2:ND | 0.8:0.8 | 99:99 | |||||||
38 | 72/F | 1/4 | 2.50/Legionnaires | 4.8:10.0 | 39.7:44.2 | 1.0:1.0 | 94:72 | |||||||
39 | 34/M | 1/14 | 2.75/trauma-fract | 13.0:5.0 | 43.7:ND | 0.7:0.4 | 209:283 |
Collection of bronchoalveolar lavage (BAL) and tracheal aspirate (ASP) fluid. As part of an unrelated study (Ethics Review Committee Permit No. 83/93), BAL was performed on 17 of the normal subjects using a flexible bronchoscope (Type P10; Olympus Optical Company, Tokyo, Japan), which was introduced transnasally with the subject supine. The tip of the bronchoscope was wedged in a right middle lobe segmental bronchus and 4 × 20-ml volumes of saline at 37° C instilled and withdrawn sequentially (12).
ASP fluid, usually from the right intermediated bronchus, was obtained during routine suctioning of the airways of the additional patients and stored at −20° C for batch analysis.
Lipids were extracted from the lyophilized BAL and ASP fluids by the method of Blight and Dyer (12, 13). Disaturated phospholipids were separated by the method of Mason and associates and the phospholipid content determined by measuring inorganic phosphorus by the Bartlett method (12, 13).
Sample treatment. Two-milliliter aliquots of BAL and ASP fluids were dialyzed against 6 L H2O for 24 h at 4° C, lyophilized, and used without further treatment. Surfactant, from pooled BAL fluid comprising equal volumes from 17 subjects, was isolated by centrifugation at 150,000 × g (max) for 1.5 h. The supernatant was filtered through a 0.2-μm Acrodysc filter (No. 4192 Sterile Acrodysc; Gelman Sciences, Ann Arbor, MI). Both the filtrate and the surfactant pellet were lyophilized. The mucoid nature of ASP fluid prevented its separation into equivalent fractions. Protein G Sepharose 4 Fast Flow (Pharmacia LKB Biotechnology, Uppsala, Sweden) bound fractions, immunoaffinity-isolated protein, N Protein Plasma Standard (human) (Behringwerke AG, Marburg, Germany) containing 1.63 mg/ml of haptoglobin, and plasma were used without dialysis or lyophilization. Unless stated otherwise, none of the samples were delipidated.
Single-dimensional electrophoresis. Samples containing ∼ 2.5 μg of protein were separated using a Laemmli buffer system with 0.75-mm-thick polyacrylamide gels (7, 13). The amount of protein loaded was determined gravimetrically except in the case of the plasma samples where protein concentration was assumed to be 76 μg/μl. Lyophilized undelipidated ASP and BAL samples were assumed to contain ∼ 10% protein. In the case of the Sepharose 4 Fast Flow bound fractions below, the protein loaded was determined by trial and error.
Two-dimensional electrophoresis. Two-dimensional electrophoretic samples were desalted and isoelectric focusing (IEF) of samples containing ∼ 20 μg of proteins was performed using an Immobiline DryStrip Kit (pH 4.0 to 7.0) (Pharmacia LKB Biotechnology) (12). Separation in the second dimension (SDS-PAGE) was performed with 15% polyacrylamide gels using a Laemmli buffer system on a vertical electrophoresis apparatus. Molecular weight and carbamylyte IEF standards were obtained from Pharmacia LKB Biotechnology.
Gels were stained with silver or the proteins transferred onto nitrocellulose membrane (7, 12, 13).
Protein blotting and immunochemical staining. Nonspecific protein binding sites were saturated by incubating the membranes with 5% bovine plasma albumin (wt/vol) (BSA, Fraction V; Sigma Chemical Co., St. Louis, MO) (7, 12, 13). The membranes were then incubated in 1% BSA (wt/vol) with either: Po-B (8 μg/ml), and antibody raised against alveolar proteinosis mature SP-B (13); 4633 (8 μg/ml), an antiserum directed against the entire recombinant human SP-B precursor protein (14, 15); 55019 (∼ 20 μg/ml), an antiserum generated against recombinant human N-terminal propeptide (residues 24 to 200) (16); 96189 (∼ 10 μg/ml), an antiserum generated against recombinant human C-terminal propeptide (residues 280 to 381) (16); or anti-human haptoglobin IgG (1.1 μg/ml) (Behringwerke) (7, 12, 13). Controls consisted of incubating equivalent blots with pre-immune serum, secondary antibody alone, or with Po-B (8 μg/ml) after first pre-incubating the antibody for 90 min with full-length recombinant SP-B (500 μg/ ml) (17). After washing, the membranes were incubated with alkaline phosphatase–conjugated sheep anti-rabbit polyclonal IgG (1 μg/ml) (Silenus Laboratories, Melbourne, Australia). The blots were developed using a Protoblot Immunoscreening System (Promega, Madison, WI). Antibodies 4633, 55019, 96189, and full-length recombinant SP-B were kindly provided by Drs. S. Lin, T. Weaver, and J. Whitsett (University of Cincinnati).
Immunoaffinity isolation. Po-B plasma immunoreactivity was further characterized in two ways: (1) Protein G Sepharose 4 Fast Flow (1 ml) was resuspended in 2 ml of 0.1 M Tris and 0.5 M NaCl containing 0.1% NaN3 (wt/vol) (pH 7.4) (Tris buffer). The sepharose was agitated for 1.5 h at room temperature with 0.5 ml of pooled plasma, composed of equal volumes from six patients with ARDS. The sepharose was removed by centrifugation and washed with 6 × 10-ml volumes of Tris buffer. The plasma was diluted to 4 ml in Tris buffer, divided, and agitated overnight at 4° C with 2.5 μl of Po-B or 2.5 μl of preimmune sera. Each sample was then mixed at room temperature with 50 μl of unused Protein G Sepharose 4 Fast Flow. After 2 h, the sepharose was isolated by centrifugation and washed with 7 × 5-ml volumes of Tris buffer. At each stage of the procedure, aliquots of the plasma were dialyzed against water. Each fraction was lyophilized and characterized by electrophoresis and immunochemical staining. (2) Po-B (1.5 ml) was coupled with 5.25 ml of CNBr-activated Sepharose 4B (Pharmacia LKB Biotechnology). The Po-B bound sepharose was suspended in 15 ml of Tris buffer containing 1 ml of the pooled ARDS plasma. After agitating at 4° C overnight, the Po-B bound sepharose was isolated onto a scintered glass funnel and washed with 10 × 20 ml volumes of Tris buffer. The IgG bound protein was desorbed with 20 ml of 0.1 M glycine, 0.15 M NaCl containing 0.1% NaN3 (wt/vol) (pH 2.3). Two milliliters of 1 M Tris (pH > 10) was added to the eluate. The eluate was dialyzed against water, lyophilized, and characterized by electrophoresis and immunochemical staining. In addition, 300 μg of the isolated protein was incubated overnight at room temperature with endoglycosidase F/N–glycosidase F (0.5 U; Sigma).
Haptoglobin levels were determined in some ARDS samples using automated immunoturbidimetry (No. 8326; Cobas Bio, Roche Diagnostica, Basel, Switzerland).
Samples were assayed in a blind, randomized manner. In order to free the SP-A and -B from any associated plasma or surfactant components, aliquots were treated with EDTA, SDS, and Triton X-100 as previously described (7). SP-B was determined using an ELISA inhibition assay similar to that described for SP-A (7). The ELISA was performed using Po-B (13). As with the SP-A ELISA, all samples were assayed in duplicate at four serial dilutions. Standards, assayed in quadruplicate, were included in each ELISA plate at 8 serial dilutions (ranging from 7.8 to 1,000 ng/ml, r > 0.98).
In order to assess the reproducibility of the procedure for the determination of SP-B, a sample was assayed on 15 separate occasions. Assay variation within a plate was assessed by assaying 16 replicates of an ARDS plasma sample.
Patient 1. A 20-yr-old male required intubation and mechanical ventilation due to fulminant hepatic failure and encephalopathy complicating Wilson's disease. His renal function (urea and creatinine) remained normal although his liver function gradually deteriorated (plasma bilirubin rose from 670 to 825 mM and the international ratio rose from 3.0 to 4.3 despite intermittent factor replacement). His chest radiograph remained normal throughout. Despite monitoring and treatment of an elevated intracranial pressure, he died 9 d after admission to the ICU as a result of cerebral edema while awaiting an emergency orthotopic liver transplant.
Patient 2. A 47-yr-old woman developed ARDS following a distal pancreatectomy performed for chronic relapsing pancreatitis. Despite intubation and ventilation for acute hypoxemic respiratory failure, her gas exchange deteriorated, requiring a Fi O2 of 0.9. A gradual improvement in gas exchange allowed the Fi O2 to be reduced to 0.5; however, an improvement in static respiratory system compliance was delayed. She was extubated 10 d after admission and made an uneventful recovery.
Patient 3. A 64-yr-old man developed acute respiratory failure following a right upper lobectomy for squamous cell carcinoma of the lung. Despite intubation and mechanical ventilation, he required a Fi O2 of 1.0 and 10 cm H2O positive end-expiratory pressure (PEEP) to achieve a PaO2 of 68 mm Hg. Although his central venous pressure and pulmonary arterial wedge pressure were not elevated, he had severe pulmonary hypertension. Inhaled nitric oxide at 10 ppm was commenced, and as attempts at discontinuation resulted in profound hypoxemia, this was continued until his death 20 d later. Nine days after development of ARDS, dexamethasone treatment (16 mg/d) was commenced in an attempt to decrease the fibroproliferative response and continued for 5 d without clinical response.
Plasma reactivity. In blots of APE and ARDS plasma separated under both reducing and nonreducing conditions, Po-B consistently reacted strongly with a peptide of apparent Mr ∼ 42 to ∼ 45 kD (Figure 1). Lesser reactivities, corresponding to peptides of relative Mr ∼ 24 to ∼ 26 kD and ∼ 45 to ∼ 65 kD, were also evident in some samples. Although essentially the same pattern of reactivity was obtained using 4633, this antibody exhibited minor reactivity with a peptide of apparent reduced Mr ∼ 80 kD.

Fig. 1. Electrophoretic (left) and immunochemical (right) staining patterns of plasma from different patients with acute cardiogenic pulmonary edema (APE) or acute respiratory distress syndrome (ARDS). The samples were separated under both reducing and nonreducing conditions on 12% polyacrylamide gels. Mr standards (std) are labeled. The electrophoretic staining pattern was obtained using silver. The immunochemical staining patterns were obtained using antibody raised against mature human SP-B (Po-B) (left blots) or the entire recombinant human SP-B precursor protein (4633) (right blots).
[More] [Minimize]Similar patterns of reactivity, albeit of lesser intensity, were observed with normal (Figure 2) and OD (not shown) plasma. Qualitatively, essentially the same patterns of reactivity were obtained using Po-B, 4633, 96189, or 55019 (Figure 2). No reactivity was detected when the blots were incubated with preimmune serum, secondary antibody alone, or Po-B after first pre-incubating the antibody for 90 min with full-length recombinant SP-B. This was also true of the blots of the plasma, BAL, and ASP fluids, described below. Antibody against haptoglobin reacted strongly with a peptide of apparent Mr 45 to 48 kD under reducing conditions and ∼ 100 and ∼ 200 kD under nonreducing conditions (not shown).

Fig. 2. Electrophoretic (top) and immunochemical (bottom) staining patterns of ARDS tracheal aspirate fluid (1), normal bronchoalveolar lavage (BAL) surfactant (2), normal BAL supernatant (3), pooled ARDS plasma (4), pooled normal plasma (5), full-length recombinant SP-B standard (∼ 50 ng) (6), and mature SP-B standard (∼ 400 μg) (7). The samples were separated under reducing conditions on 12% polyacrylamide gels. Mr standards (std) are labeled. The electrophoretic staining pattern was obtained using silver. The immunochemical staining patterns were obtained using antibody raised against mature human SP-B (Po-B), the entire recombinant human SP-B precursor protein (4633), recombinant human C-terminal propeptide SP-B (residues 280 to 381) (96189), and recombinant human N-terminal propeptide SP-B (residues 24 to 200) (55019). Legend as in Figure 1.
[More] [Minimize]BAL and ASP fluid reactivity. Normal BAL fluid. Although Po-B reacted with the mature SP-B standard separated under reducing conditions (Figure 2), little reactivity corresponding to the mature protein was evident in normal BAL fluid. As expected, no reactivity was detected when 4633, 96189, or 55019 were used. When normal BAL fluid was separated into a lipophilic surfactant pellet and a supernatant, reactivity corresponding to mature SP-B was predominantly associated with the supernatant. Despite this, ELISA analysis in the presence of EDTA and detergent indicated that the pellet in fact contained ∼ 10 times more immunoreactive SP-B than the supernatant. The pellet also contained ∼ 20 times more disaturated phospholipids than the supernatant (data not shown). Possibly, mature SP-B epitopes remain hidden in the presence of surfactant lipids (10). Alternatively, high concentrations of lipids may prevent efficient transfer of mature SP-B onto nitrocellulose membrane.
Whereas the major peptide detected in the plasma blots had a relative Mr ∼ 42 to ∼ 45 kD, in normal BAL fluid reactivity corresponding to a peptide of apparent Mr ∼ 24 to ∼ 26 kD was more abundant. This was true of samples separated under both reducing (Figure 2) and nonreducing conditions (not shown). The peptides of relative Mr ∼ 24 to ∼ 26 kD and ∼ 42 to ∼ 45 kD were predominantly associated with the supernatant rather than the surfactant pellet, and whereas the antibodies Po-B, 4633, and 96189 all exhibited strong reactivity with both, 55019 only exhibited appreciable reactivity with the peptide of relative Mr ∼ 42 to ∼ 45 kD.
ARDS ASP fluid. As with normal BAL fluid, little reactivity corresponding to mature SP-B was evident in ARDS ASP fluid. Whereas, qualitatively, most reactivities were common to both BAL and ASP fluid, all SP-B antibodies exhibited minor reactivity towards a peptide of apparent reduced Mr ∼ 80 kD (Figure 2). This reactivity, which was also apparent in blots of OD ASP fluid, was never observed under nonreducing conditions (not shown) or prevalent in normal BAL fluid. Because all four antibodies recognized the peptide, it is not likely to be an oligomer of mature SP-B. High Mr immunoreactive SP-B forms have been observed by Haagsman and associates (personal communication) and others (9). Possibly, nonreducible forms arise in ventilated patients through reduced glutathione levels or increases in oxidative enzymes, together with impairment of the mucociliary escalator.
In contrast to the consistent immunostaining pattern obtained from blots of normal BAL fluid, quantitatively, the pattern of reactivity obtained with ARDS ASP fluid varied markedly between patients, as well as daily during a patient's stay in the ICU (Figure 3). Two additional reactivities corresponding to peptides of relative Mr ∼ 18 and ∼ 23 kD were detected in some normal BAL and ARDS ASP fluids when 4633, but not Po-B, was used (Figure 3). The ∼ 18 kD peptide was occasionally detected by 96189 but not by 55019, whereas the ∼ 23 kD peptide was occasionally detected by 55019 but not by 96189. Possibly, these reactivities correspond to the cleaved amino- and carboxy-terminus of SP-B proprotein.

Fig. 3. Electrophoretic (left) and immunochemical (right) staining patterns of BAL fluid from different normal subjects and of consecutive daily tracheal aspirate (ASP) fluids from two typical ARDS patients (1, 2). The electrophoretic staining pattern was obtained using silver. The samples were separated under reducing conditions on 15% polyacrylamide gels. Legend as in Figure 1.
[More] [Minimize]Immunoprecipitation and affinity isolation from ARDS plasma. The Po-B and 4633 immunochemical staining pattern of the ARDS plasma used for the absorption studies was similar to that described above, with the major reactivity corresponding to a peptide of apparent Mr ∼ 42 to ∼ 45 kD (Figure 4). The pattern of reactivity was not altered by absorbing out the native IgG or by incubating the native IgG free plasma with rabbit Po-B preimmune serum or Po-B. Neither Po-B nor 4633 exhibited appreciable reactivity with the native IgG or the rabbit IgG isolated after incubating the plasma with pre-immune serum.

Fig. 4. Electrophoretic (left) and immunochemical (right) characterization of protein immunoprecipitated or affinity-isolated, using antibody raised against mature human SP-B, from plasma pooled from equal volumes from six patients with ARDS. The samples were separated under reducing conditions on 15% polyacrylamide gels. Lane 1: whole pooled ARDS plasma; lane 2: pooled ARDS plasma desorbed of native IgG; lane 3: pooled ARDS plasma desorbed of native IgG but with rabbit pre-immune serum; lane 4: pooled ARDS plasma desorbed of native IgG but with antibody Po-B; lane 5: desorbed native IgG; lane 6: desorbed rabbit pre-immune serum (IgG); lane 7: immunoaffinity-isolated protein: lanes 8 and 10: immunoprecipitated protein (at two different loads); lanes 9 and 11: immunoaffinity-isolated protein (at two different loads) after incubating overnight at room temperature with endoglycosidase F/N–glycosidase F; lane 12: mature SP-B standard. The electrophoretic staining pattern was obtained using silver. Legend as in Figure 1.
[More] [Minimize]The silver and immunochemical staining pattern of protein isolated by immunoprecipitation was similar to that isolated by affinity isolation. Under reducing (Figure 4) and nonreducing (not shown) conditions, the major protein had an apparent Mr ∼ 42 to ∼ 45 kD. Lesser amounts of protein with Mr corresponding to ∼ 22 to ∼ 28 kD and ∼ 45 to ∼ 65 kD were also evident. No change in Mr was observed when the affinity-isolated peptides were treated with endoglycosidase.
Two-dimensional electrophoresis. Protein isolated by immunoaffinity absorption separated under reducing conditions as a major series of five spots spanning a Mr range from ∼ 42 to ∼ 47 kD and a pI range from 5.19 to 6.19 (Figure 5). In addition, a second single major spot that corresponded to Mr ∼ 24 kD and a pI of 6.40 was also evident. Two minor peptides (Mr ∼ 64 kD, pI 6.00 and Mr ∼ 69 kD, pI 6.56) were identified as antithrombin III and albumin by reference to two-dimensional electrophoretic profiles of human plasma (18). Because neither Po-B nor 4633 reacted with these peptides in the plasma blots described above, it is likely that both are residual contaminants, nonspecifically associated with the affinity media during the isolation procedure.

Fig. 5. The two-dimensional SDS-PAGE pattern of protein isolated from pooled ARDS plasma by immunoaffinity adsorption using antibody raised against mature human SP-B (peptides lane 7 in Figure 4). The sample was separated under reducing conditions on a 15% polyacrylamide gel. Mr standards are labeled. The relative Mr and isoelectric points of the major peptides are shown. The electrophoretic staining pattern was obtained using silver.
[More] [Minimize]Statistics. We have tested both the SP-A and -B data from the ventilated patients for normality using the Sharpiro-Wilk test. The distribution of both variables was clearly nonlinear, with correlations of r = 0.90 for SP-A and r = 0.85 for SP-B. Since this leads to a rejection of the null hypothesis of normality, we have used nonparametric statistics. Unless stated otherwise, results are expressed as median (range). The Mann-Whitney U test or the Wilcoxon matched pairs sign rank test was used for all comparisons. For comparisons with n > 20, significance using the Mann-Whitney U test was determined from the normal deviate Z with reference to normal frequency distribution tables. The association between measured variables was tested using the Spearman rank order correlation test. We have previously reported the concentration of SP-A in some of the normal subjects of OD, APE, and ARDS patients (7). Because inclusion of the additional subjects did not alter our previous findings, we now present only the SP-B and SP-B/SP-A data. The reproducibility of the SP-B assay is illustrated by the coefficient of variation (CV). The ELISA had an intraassay CV of 5.96% (n = 16) and an interassay CV of 8.64% (n = 15).
Normal subjects and OD patients. Plasma SP-B in normal subjects (1,696 ng/ml [1,197 to 2,550], n = 33) was not significantly different from that in the OD patients (Table 2). The plasma SP-B/SP-A ratio of normal subjects (9.8 [5.3 to 19.8]) was not significantly different from that in OD patients (Table 2).
Patient | SP-B (ng/ml ) | SP-B/SP-A | Outcome | |||
---|---|---|---|---|---|---|
OD | 1,894 | 15.0 | S | |||
1,960 | 7.7 | S | ||||
2,548 | 15.2 | S | ||||
1,390 | 10.4 | S | ||||
1,805 | 10.6 | S | ||||
1,059 | 5.8 | S | ||||
2,144:2,241 | 10.4:11.3 | DNC* | ||||
APE | 8,408:8,070 | 28.6:20.2 | S | |||
2,614:3,331 | 9.9:13.0 | S | ||||
2,984:4,167 | 16.1:20.0 | S | ||||
3,973:3,632 | 12.0:11.2 | S | ||||
2,288:4,820 | 9.7:18.0 | S | ||||
3,244:3,200 | 9.0:9.1 | S | ||||
4,933:6,005 | 23.5:33.0 | S | ||||
2,027:2,597 | 11.3:12.7 | S | ||||
1,425:1,727 | 6.3:6.8 | S | ||||
4,560:3,407 | 12.5:9.4 | S | ||||
ARDS | 35,526:39,786 | 25.8:42.3 | DA | |||
3,505 | 13.6 | DNC | ||||
15,831 | 29.9 | DNC | ||||
12,966:5,600 | 13.4:9.6 | S† | ||||
4,010:3,910 | 5.7:6.4 | DA | ||||
4,708:6,621 | 9.1:21.9 | DNC | ||||
1,648:1,363 | 7.7:8.0 | S | ||||
10,608:3,058 | 29.0:2.5 | DA | ||||
12,576:8,108 | 22.6:10.0 | DA | ||||
5,206:1,926 | 13.8:5.1 | S | ||||
1,876:1,718 | 5.1:6.3 | DNC | ||||
2,159:1,380 | 9.6:3.1 | DNC | ||||
5,635:6,821 | 16.7:14.1 | DA | ||||
11,832:17,022 | 17.3:13.6 | DA | ||||
8,313:35,436 | 24.9:35.2 | DA | ||||
3,369:2,455 | 7.3:11.5 | DNC | ||||
8,460:13,242 | 35.5:24.1 | S | ||||
3,033:3,636 | 14.2:10.7 | S | ||||
4,812:13,297 | 13.2:42.5 | DA‡ | ||||
9,627:9,949 | 26.3:24.2 | DNC | ||||
13,412:4,413 | 25.9:5.6 | DA | ||||
11,402:4,301 | 14.2:11.5 | S |
APE patients. Plasma SP-B in APE patients at the time of entry into the study was greater than in normal subjects (p < 0.0001) or OD patients (p < 0.01) (Table 2). After 24 h of mask continuous positive airway pressure (CPAP), the lung function of the APE patients had clearly improved, as reflected by the changes in PaO2 /Fi O2 , Fi O2 , and PEEP (7). Despite this, plasma SP-B did not change (Table 2). The plasma SP-B/SP-A ratio in APE patients at the time of their entry into the study was not significantly greater than in normal subjects or OD patients and was not altered after 24 h of mask CPAP (Table 2).
ARDS patients. No relationship was observed between plasma levels of haptoglobin and SP-B. Plasma SP-B in the ARDS patients on Day 1 was greater than in the normal subjects (p < 0.0001), the OD patients (p < 0.001), and the APE patients (p < 0.025) (Table 2). Plasma SP-B levels subsequently varied markedly from one day to the next in a manner that mostly reflected lung function, as described below. However, plasma SP-B, but not SP-A, tended to fall during the patient's stay in the ICU such that the levels on Day 1 (Table 2) were almost twice as high as those on all subsequent days (3,636 ng/ml [961 to 45,164], n = 203) (p < 0.005). These reduced concentrations were no different from the levels in the APE patients but remained greater than in the normal subjects (p < 0.0001) and OD patients (p < 0.005). As distinct from the other groups, plasma SP-A and -B were related in the ARDS patients, both on Day 1 (p < 0.0001, rs = 0.69, n = 22) and during their stay in the ICU (p < 0.0001, rs = 0.31, n = 203).
The plasma SP-B/SP-A ratio in the ARDS patients on Day 1 was greater than in the normal subjects (p < 0.005) and the OD patients (p < 0.05) (Table 2) but no different from that in the APE patients. Since plasma SP-B, but not SP-A, tended to fall during the patient's stay in the ICU, the SP-B/SP-A ratio (10.0 [2.5 to 86.0], n = 203) was also less (p < 0.01) during the patient's admission to the ICU than on Day 1 (Table 2). The reduced ratio did not differ from that in the normal subjects or the OD and APE patients.
In patients who had acquired the syndrome as the result of a direct insult (via the airways), plasma SP-B, but not SP-A, was greater than in patients who had developed ARDS indirectly (via the circulation). This was true both on Day 1 (p < 0.05) (Table 2) and during the patients' stay in the ICU (p < 0.0001) (direct: SP-B = 6,051 ng/ml [2,458 to 35,436], n = 62; indirect: SP-B = 2,661 ng/ml [961 to 45,164], n = 141). Consequently, the plasma SP-B/SP-A ratio was also greater in the direct ARDS patients than in the indirect ARDS patients, again, both on Day 1 (p < 0.01) (Table 2) and during their stay in the ICU (p < 0.0001) (direct: SP-B/SP-A = 15.9 [2.5 to 86.0], indirect: 8.3 [2.5 to 49.6]). Although the patients who acquired the syndrome directly also tended to have the worst blood oxygenation and ΔV/ΔP, analysis of covariance suggested that the differences in plasma SP-B and the SP-B/SP-A ratio were not related to severity.
Outcome. Circulating levels of SP-B on Day 1 tended to be greater in patients who subsequently died from ARDS than in those who survived or died from non–ARDS-related causes (p < 0.05 in both cases; Table 2). In patients who survived, plasma SP-B did not change during their time in the ICU (6,537 ng/ml [1,294 to 45,164], n = 54) compared with their levels on Day 1 (Table 2), whereas the levels tended to fall by ∼ 50% in patients who subsequently died from ARDS (5,263 ng/ml [2,458 to 39,786], p < 0.025, n = 64) or nonARDS-related causes (2,202 ng/ml [961 to 9,949], p < 0.05, n = 85). This was in marked contrast to the levels of SP-A, which decreased in patients who survived, remained the same in those who died from non–ARDS-related causes, and increased in those who died from ARDS (7).
PCP. On Day 1, PCP in the ARDS patients was 21.6 mm Hg (11.2 to 29.6, n = 22) and did not change in any of the patients during their stay in the ICU. Although PCP was not related to plasma SP-A or -B, the ratio of SP-B/SP-A was directly related to PCP during the patient's stay in the ICU (p < 0.05, rs = 0.17, n = 144).
No clinical measurements were performed in the normal group and we have assumed PaO2 /Fi O2 to be 500 and ΔV/ΔP to be 100 ml/cm H2O for these subjects. ΔV/ΔP was not determined in the APE patients.
Blood oxygenation (PaO2 / Fi O2 ). On Day 1, plasma SP-B was inversely related to PaO2 /Fi O2 , with (p < 0.0001, rs = −0.74, n = 72) (Figure 6) and without (p < 0.005, rs = −0.48, n = 39) inclusion of the normal subjects. In the ventilated patients, the relationship persisted in the sequential samples collected during their stay in the ICU (p < 0.0001, rs = −0.36, n = 260 [OD: n = 15; APE: n = 20; ARDS: n = 225]). The relationship also remained within the ARDS patients during their time in the ICU (p < 0.0001, rs = −0.34, n = 225). The plasma SP-B/SP-A ratio was also inversely related to PaO2 /Fi O2 in the normal subjects and ventilated patients on Day 1 (p < 0.01, rs = −0.30, n = 72) (Figure 1). In the ventilated patients, the relationship persisted in the sequential samples (p < 0.025, rs = −0.14, n = 260). The relationship also remained within the ARDS patients during their time in the ICU (p < 0.001, rs = −0.22, n = 225).

Fig. 6. The relationship between plasma immunoreactive SP-B (p < 0.0001, rs = −0.74, n = 72) and the plasma SP-B/SP-A ratio (p < 0.010, rs = −0.30, n = 72) with PaO2 /Fi O2 . The samples were obtained from normal subjects (closed squares) and from the following groups of ventilated patients: no evidence of cardiorespiratory disease (closed circles), acute cardiogenic pulmonary edema (open squares), and ARDS (open circles).
[More] [Minimize]Static respiratory system compliance (ΔV/ΔP). On Day 1, plasma SP-B was inversely related to ΔV/ΔP, with (p < 0.0001, rs = −0.74, n = 57) and without (p < 0.005, rs = −0.58, n = 24) inclusion of the normal subjects. In the ventilated patients, the relationship persisted in the sequential samples collected during their stay in the ICU (p < 0.0001, rs = −0.46, n = 168 [OD: n = 7; ARDS: n = 161]). The relationship also remained within the ARDS patients, both on Day 1 (p < 0.05, rs = −0.49, n = 22) and during the course of their admission (p < 0.0001, rs = −0.44, n = 161). The plasma SP-B/SP-A ratio was also inversely related to ΔV/ΔP in the normal subjects and ventilated patients on Day 1 (p < 0.005, rs = −0.37, n = 57). In the ventilated patients, the relationship persisted in the sequential samples (p < 0.01, rs = −0.21, n = 168). The relationship also remained within the ARDS patients during their time in the ICU (p < 0.005, rs = −0.23, n = 161).
In the 11 matching samples, there was a small but significant (p < 0.005) difference in the SP-B concentration in plasma prepared from arterial blood (4,708 ng/ml [2,390 to 17,022]) and from mixed venous blood (3,706 ng/ml [1,763 to 15,258]). The arterial/venous plasma SP-B ratio was 1.2 (1.0 to 2.2). If we assume an average blood volume of 5 L, with 85% of the blood residing in the systemic circulation, then from the patients' cardiac outputs (not shown) we estimate that the half-life of plasma immunoreactive SP-B is 153 s (51 to 390 s, n = 9) (two matching pairs with arterial/venous plasma SP-B ratios of 1.0 were excluded for this estimate), less than we have previously estimated for serum SP-A (7). Consistent with this finding, the SP-B/SP-A ratio was also less (p < 0.05) in plasma prepared from arterial blood (9.8 [3.5 to 24.9]) than that prepared from mixed venous blood (10.6 [3.4 to 35.5]).
Patient 1. Despite hepatic failure and encephalopathy complicating Wilson's disease, lung function of this patient remained relatively stable during his stay in the ICU (Figure 7, left column). During this time, plasma SP-B and the SP-B/SP-A ratio remained low and relatively constant.

Fig. 7. The daily variation in plasma SP-A (open circles, top row), SP-B (open circles, middle row), and the SP-B/SP-A ratio (open circles, bottom row) and matching measures of PaO2 /Fi O2 (closed squares) and static respiratory system compliance (ΔV/ΔP) (closed circles) obtained from the three cases presented in the text (columns 1 through 3). The variables have been presented as changes about the mean determined during the patient's entry in the study. Legend as in Figure 1.
[More] [Minimize]Patient 2. The low ΔV/ΔP and PaO2 at the time of admission to the ICU was accompanied by greatly elevated plasma SP-A and -B (Figure 7, middle column). Whereas PaO2 /Fi O2 remained relatively constant, ΔV/ΔP worsened over the next 3 d. During this period, plasma SP-B increased in a concomitant manner. After Day 4, there was a gradual improvement in both PaO2 /Fi O2 and ΔV/ΔP, accompanied by a progressive decrease in plasma SP-A and -B. SP-A and -B levels were inversely related to PaO2 /Fi O2 (p < 0.025, rs = −0.65; p < 0.025, rs = −0.70; n = 9; respectively). SP-B, but not SP-A, was also inversely related to ΔV/ΔP (p < 0.001, rs = −0.93, n = 8). The SP-B/SP-A ratio was also inversely related to PaO2 /Fi O2 (p < 0.025, rs = −0.72, n = 9) and ΔV/ΔP (p < 0.005, rs = −0.86, n = 8).
Patient 3. The low PaO2 /Fi O2 and reduced ΔV/ΔP at the time of diagnosis of ARDS was accompanied by plasma SP-A and -B levels ∼ 2-fold and ∼ 4-fold greater, respectively, than observed in normal subjects (Figure 7, right column). Over the next 3 wk, PaO2 /Fi O2 remained depressed but relatively constant. In contrast, ΔV/ΔP varied almost daily. These changes were reflected in concomitant variations in plasma SP-A and -B. SP-A and -B levels were inversely related to ΔV/ΔP (p < 0.05, rs = −0.48; p < 0.05, rs = −0.48; n = 18; respectively). The SP-B/SP-A ratio was almost inversely related to ΔV/ΔP (p < 0.025, rs = −0.52).
We have found that immunoreactive SP-A (7) and -B breach the alveolocapillary barrier in a manner inversely related to lung function. Furthermore, the ratio of plasma SP-B/SP-A is also inversely related to both blood oxygenation and static respiratory system compliance, suggesting that SP-B breaches the alveolocapillary barrier more readily than SP-A and so may provide a more sensitive marker of lung injury.
We have purified mature human SP-B and raised a polyclonal antibody that reacts strongly with the mature protein under both reducing and nonreducing conditions. We have immunoabsorbed our antibody against serum in order to remove any nonspecific circulating determinants present and have compared its specificity with antibodies raised by others, including 4633 (raised against a synthetic fusion protein of the human full-length SP-B cDNA (14, 15), 96189 (generated against recombinant human C-terminal propeptide, residues 280 to 381) (16), and 55019 (generated against recombinant human N-terminal propeptide, residues 24 to 200) (16). Importantly, in terms of characterizing our protein blots, 4633, 96189, and 55019 do not react with mature SP-B (14, 15).
The hydrophobic mature SP-B is first synthesized as a precursor, which, after signal peptide cleavage, probably in the endoplasmic reticulum after translocation, and glycosylation, yields a hydrophilic proprotein (Mr ∼ 42 kD). It is generally accepted that processing of the proprotein involves at least two distinct proteolytic events: first, cleavage of 200 amino acids at the N-terminus to afford a ∼ 25 kD intermediate and, second, cleavage of 102 amino acids at the C-terminus to yield the mature protein. Lamellar bodies, the major secretory source of surfactant lipids, are associated with mature SP-B but little, if any, of the proprotein or ∼ 25 kD intermediate (11, 19). However, numerous quandaries remain regarding SP-B secretion and processing.
Pulse-chase labeling studies performed by Weaver and Whitsett in rat isolated type II cell preparations indicate that SP-B proprotein is primarily and rapidly secreted, presumably constitutively, and is then cleaved first at the amino terminus to afford the ∼ 25 kD intermediate (11). Mature SP-B was also detected in the media; however, it is unclear whether the mature peptide associated with lamellar bodies is related to the intracellular processing and secretion of the protein or whether it associates with lamellar bodies only following extracellular processing and reuptake. Voorhout and associates (19) suggest that the precursor form is proteolytically processed to the mature peptide between the Golgi complex and lamellar bodies, probably in multivesicular bodies, and conclude that extracellular processing of SP-B is unlikely. However, although we have recently shown that lamellar bodies are indeed enriched in immunoreactive SP-B (13), we also found that the immunoreactive SP-B/phospholipid ratio in the alveolus is more than an order of magnitude greater, leading us to conclude that, in the absence of differential concentration-dependent clearance of SP-B and phospholipids, SP-B in the alveolus must be partially derived from material secreted independent of lamellar bodies.
Despite the common assertion that only mature SP-B is present in BAL fluid, we are not aware of any reports of immuno-analysis of whole lung lavage. Our present findings provide clear evidence that the presence of extracellular SP-B proprotein and intermediate is not just an artifact of the in vitro model previously used (11). The major reactively detected by Po-B, 4633, and 96189 in whole, nondelipidated, normal BAL and ARDS ASP fluid corresponded to a peptide of apparent Mr ∼ 24 to ∼ 26 kD under both reducing and nonreducing conditions. Because this peptide did not react with 55019, it is probably processing intermediate (11, 14, 16). Lesser reactivities, corresponding to peptides with nominal Mr ∼ 8 kD and ∼ 42 to ∼ 45 kD under reducing conditions and ∼ 18 kD and ∼ 42 to ∼ 45 kD under nonreducing conditions, were also detected. Since the larger peptide was responsive to all four antibodies, whereas the smaller reducible peptide only reacted with Po-B, we suggest that the peptides are proprotein and mature SP-B, respectively. Peptides possibly corresponding to the cleaved amino- and carboxyl-terminus of the proprotein were also occasionally detected and identified in the same manner. Importantly, when whole BAL was ultracentrifuged and filtered, the proprotein and processing intermediate were clearly predominantly associated with the BAL supernatant rather than the lipophilic surfactant pellet. Because of the different titer of the antibodies used, we did not attempt to quantify the relative amounts of proprotein, processing intermediate, and mature SP-B. This task is made more difficult by the extreme hydrophobicity of mature SP-B and its affinity for surfactant lipids.
In contrast to SP-A, which our studies (7) and those of Kuroki and coworkers (2) suggest may exist in the blood as a complex with immunoglobulins G and M, we found no evidence that plasma immunoreactive SP-B is associated with other proteins. Rather, the immunochemical staining pattern was similar to that observed in ASP and BAL fluid, except that the reactivity corresponding to a nominal Mr ∼ 42 to ∼ 45 kD rather than ∼ 24 to ∼ 26 kD, predominated. The broad electrophoretic staining pattern of both peptides is consistent with glycosylation. In the case of the peptide with the nominal Mr ∼ 42 to 45 kD, the charge train seen by two-dimensional electrophoresis is consistent with sialyation of the oligosaccharide tree. Endoglycosidase treatment had no apparent effect on the Mr of either peptide, probably suggesting that the treatment did not work. Possibly, the differential proportions of SP-B proprotein and processing intermediate in the alveolus and in plasma reflect differences in hydrophobicity. The two-dimensional electrophoretic coordinates of the affinity-purified plasma material closely agrees with that of SP-B proprotein and processing intermediate (14, 15). Although the β-chain of haptoglobin (∼ 45 kD), an acute-phase protein that may be elevated in inflammatory diseases, has similar two-dimensional electrophoretic characteristics to the proprotein (18), our single-dimensional immunochemical staining pattern of both plasma and the affinity-isolated protein was similar under both reducing and nonreducing conditions. Notably, no reactivity indicative of native haptoglobin (∼ 100 kD and oligomers > 200 kD) (20) was observed under nonreducing conditions. Moreover, we observed no cross-reactivity between an antibody against haptoglobin and the protein isolated by affinity absorption.
Our analysis suggests that plasma contains very little mature SP-B. Since mature SP-B is normally intimately associated with alveolar surfactant phospholipids, it is possible that complexes so formed are too large to breach the alveolocapillary barrier. Alternatively, it may be undetectable on protein blots or mature SP-B may breach the lung barrier but is then rapidly removed from the circulation or buried in membranes.
We do not believe that our plasma SP-B determinations are influenced by nonspecific reactivities. It is difficult to conceive that the strong correlation we observed between plasma SP-A and -B in our ARDS patients is fortuitous, given that the proteins share few structural similarities. We also believe that the close reciprocal matching in the levels of plasma SP-B with the changes in static respiratory system compliance and blood oxygenation in the two ARDS cases presented indicate that the data are robust.
Plasma SP-A and -B in normal subjects. Surprisingly, our results indicate that immunoreactive SP-A and -B are normally present in the systemic circulation. Leakage of Clara cell secretory protein (CC10) into the bloodstream of normal subjects has also been reported by Bernard and associates (21) and Nomori and colleagues (22), who suggest that circulating CC10 is derived exclusively from the respiratory tract. By way of comparison, Bernard and associates (21) report that the mean CC10 concentration in normal serum is ∼ 82 ng/ml whereas in BAL fluid it is ∼ 3.6 μg/ml, when lavaging with 125 ml of saline. The equivalent measures of immunoreactive SP-B in our normal subjects were ∼ 1,650 ng/ml and ∼ 11.5 μg/ml when lavaging with 80 ml (unpublished findings), respectively. Therefore, all other things being equal, normal circulating levels of immunoreactive SP-B are ∼ 6.4 times relatively higher than those of CC10. This may reflect a number of factors, including that CC10 is mostly expressed in the nonciliated columnar epithelial cells in the lower and upper airways (22) rather than in the alveolus, where permeability and surface area are greatest (6, 23). Moreover, protein transport from the airways across bronchial epithelium is usually accompanied by protein degradation whereas proteins cross the alveolar epithelium largely intact (6).
Relative SP-A and -B levels in lung and blood. We have found that the SP-B/SP-A ratio in normal BAL fluid is 1.72 ± 0.12 (mean ± SEM, n = 68 from 17 subjects), whereas in ARDS ASP fluid it is 1.67 ± 0.17 (n = 39) (unpublished findings). These contrast with those of Gregory and associates (24), which indicated a SP-B/SP-A ratio of ∼ 0.01 in normal BAL fluid and ∼ 0.02 in BAL fluid from ARDS patients, far short of the relative amount of SP-B normally associated with surfactant lipids. Possibly, the differences relate to the fact that Gregory and associates performed their analyses using an ELISA based on bovine SP-B, which may have underestimated the human SP-B content. In addition, we routinely pretreat our samples to dissociate the proteins from interfering surfactant associated lipids (7, 12, 13). Also, Gregory and associates only measured the SP-B that sedimented with surfactant lipids following high-speed centrifugation of BAL fluid. Presumably this would have excluded the water-soluble higher-Mr immunoreactive precursor forms present. On the other hand, our analyses are in good agreement with those of LeVine and coworkers (25), who recently reported that the immunoreactive SP-B/SP-A ratio is ∼ 1.6 in normal ASP fluid and ∼ 2.3 in ARDS ASP fluid. Similarly, Lesur and colleagues reported that the ratio is ∼ 1.6 in whole sheep lavage, increasing to ∼ 2.3 following acute exposure to silica (26).
The ratio of circulating SP-B/SP-A in normal subjects was ∼ 6 times greater than in the BAL fluid, whereas in ARDS patients the plasma ratio was ∼ 7 times more than ASP fluid (unpublished findings). Assuming that all of the plasma immunoreactive SP-A and -B is derived from the airspaces, then it appears that SP-B breaches the alveolocapillary barrier more readily than SP-A.
In part, this will depend not only on the relative sizes of the proteins and lung permeability but also the form available to breach the membrane barrier. SP-A binds phospholipid avidly to the extent that estimates suggest that there is little of it free in alveolar fluid (27). In contrast, we have found that the predominant forms of alveolar immunoreactive SP-B, proprotein and processing intermediate, are not bound to surfactant lipids, possibly allowing freer entry into the circulation. Consistent with this, Honda and associates (5) have observed that serum levels of SP-D are much higher than SP-A in patients with various lung diseases, the opposite of normal alveolar levels. Unlike SP-A, SP-D is also predominantly associated with the hydrophilic lavage fraction.
Protein sieving. Plasma proteins normally reach the alveolus by a bidirectional protein flux that restricts the passage of large molecules in a manner related logarithmically to the Mr of the proteins (6, 23, 28). Molecular sieving of surfactant proteins traversing the alveolocapillary barrier may explain the increased SP-B/SP-A ratio in plasma relative to that in BAL and ASP fluid. If indeed immunoreactive SP-A and -B pass directly from the alveolar hypophase into the circulation, as appears to be the case for most of the plasma proteins traversing in the opposite direction (23), we might expect that increased lung permeability would be associated with an increased flux but a loss of size selectivity, and consequently, reduced plasma SP-B/SP-A ratio. In contrast, the plasma SP-B/SP-A ratio in our ARDS patients on Day 1 was similar to that in APE patients and greater than in normal subjects and OD patients. During the patient's stay in the ICU, the absolute amounts of plasma SP-A and -B generally remained elevated, suggesting a continuing increased flux. However, while the ratio tended to decrease, it did not fall below the levels in normal subjects or OD patients, suggesting that size selectivity was not impaired.
Although the tight junctions of the epithelium, rather than the endothelium, have been regarded as the major barrier to the movement of proteins into the alveolus, it has recently been shown that the permeability of the epithelium is dynamic and regulated depending on the physiologic requirements (23). The relative degrees to which the epithelium and endothelium are damaged in ARDS appear to be related to the initiating injury (6, 23), and the work of Blake suggests that ARDS increases the radius of intermediate-sized pores rather than irreversibly destroying the membrane barrier and its sieving properties (29).
Differential entry. Possibly, SP-A and -B enter the circulation via distinct pathways that are altered to different extents in respiratory disease. Indeed, Plasma SP-B, but not SP-A, was greater in patients who had developed ARDS as the result of a direct insult, such as aspiration of gastric contents or pneumonia, than in patients who had developed ARDS indirectly. Direct denuding of alveolar epithelial type I cells may preferentially facilitate translocation of SP-B between the alveolar hypophase and the interstitium. Differential entry of SP-A and -B into the circulation may also occur via vesicular transport through type I cells. Plasmalemmal vesicles, with an average diameter of 70 nm, comprise ∼ 2% of type I cell cytoplasmic volume and may be too small for SP-A associated with surfactant lipids (23, 30). Although SP-A and -B may enter the circulation via the lymphatics, either through release into the interstitium from damaged type II cells or by differential secretion through the basement membrane (31), Honda and associates (5) have shown a remarkable fall in serum SP-D levels in proteinosis patients after whole lung lavage, consistent with direct leakage from the alveolus.
An important distinction in considering protein flux across the alveolocapillary barrier is that whereas plasma proteins may approach a bidirectional equilibrium in ARDS patients (6, 23, 28), the flux of surfactant proteins must be unidirectional. Not only is there a huge disparity between estimates of the high surfactant protein concentrations in the hypophase (27) and the amounts we detect in the circulation, but the arterial/mixed venous blood differences and rapid changes in blood levels indicate that SP-A and -B are rapidly cleared from the circulation. However, our finding that the half-life of plasma immunoreactive SP-B in ARDS patients is ∼ 2.5 min compared with ∼ 4.5 min for immunoreactive SP-A (7), suggests that, at least in so far as these patients are concerned, differential clearance cannot contribute to the higher SP-B/SP-A ratios in plasma relative to those in ASP fluid.
We do not know how SP-A and -B are cleared from blood. However, the finding that plasma levels did not increase in the OD case presented, despite hepatic failure, suggests that the liver is not solely involved. Consistent with this, there was no indication that the elevated levels in ARDS patients were related to liver function. The filtration of proteins across the renal glomerular-capillary membrane occurs with an approximate log-log relationship to Mr. Although this is also influenced by steric considerations, most proteins smaller than 50 kD pass relatively freely through the glomerulus. Virtually all (95 to 99%) filtered protein is taken up and catabolized by fusion with lysosomes in the heterolysosomes of cells in the proximal tubules (32). Therefore, the detection of immunoreactive SP-A and -B in ARDS urine (unpublished findings) suggest that glomerular filtration may normally be a major route of systemic clearance. The relative sizes of the proteins may explain the differences in their half-lives. Certainly, the rate of renal blood flow (∼ 1.2 L/min) is consistent with the rapid clearance. Indeed, many low-Mr proteins, including CC10, are freely eliminated by glomerular filtration followed by reabsorption and catabolism (33). We found no relationship between the levels of plasma immunoreactive SP-A or -B and creatinine, consistent with the findings of Kabanda and colleagues showing that in the absence of diuresis proteins normally cleared by the kidney are rapidly cleared by other unknown pathways such that serum levels remain constant rather than becoming elevated (33).
Since some insults may initially increase surfactant pool sizes indirectly as a result of a proliferation of type II cells, by causing hyperventilation, or via a stress-induced β-adrenergic stimulus, it is possible that the elevated plasma SP-A and -B in our critically ill patients reflect an upregulation in synthesis rather than increases in alveolocapillary permeability. However, our patients were in the exudative phase of the syndrome when they entered the study. We have previously argued (7) that it is doubtful that upregulation of surfactant would persist into this phase given the alveolocapillary damage present by this time. Indeed, ARDS BAL fluid has been purported to contain reduced SP-A and -B (24). Importantly, we have shown that both acute exercise and training increase the surfactant pool size, including the amounts of SP-A and -B, but has no effect on circulating levels of the proteins (12, 34). Finally, we (35) and others (36) have found no relationship between plasma SP-A and -B and levels in ASP (35) and BAL (36) fluid in normal subjects and ARDS patients. This suggests that increased alveolar SP-A and -B, per se, do not lead to elevations in circulating levels.
As we have indicated before (7), our studies suggest that leakage into the circulation must be a factor accounting for the depressed concentrations of SP-A and -B in ARDS BAL fluid (24). Assuming surfactant pool sizes are 10 to 15 mg/kg body weight, 10% of which is protein, and that SP-A and -B comprise 25 and 50% of the total immunoreactive surfactant protein pool (27), respectively, then normal adult lungs contain ∼ 20 to 30 mg of SP-A and ∼ 40 to 60 mg of immunoreactive SP-B. The plasma SP-A and -B levels we detect in our ARDS patients therefore represent ∼ 10 and ∼ 50% of the total alveolar SP-A and -B content, respectively.
We have observed highly significant inverse relationships between plasma SP-A and -B and lung function, as measured by blood oxygenation and static respiratory system compliance. Since there was no direct relationship between plasma immunoreactive SP-A or -B and PCP in ARDS patients, we believe that the leakage into the circulation most likely reflects a change in alveolocapillary permeability. Similar relationships between lung permeability and function have been observed in preterm ventilated lambs (37) and in an animal model of respiratory failure (38). Perhaps, more importantly, our work illustrates that there is a differential change in plasma SP-A and -B with lung injury, such that the ratio of SP-B/SP-A is also inversely related to lung function. Given the variability in the amount of alveolar SP-A and -B in ARDS patients (24), it is little wonder that there is a degree of scatter in the relationships we have observed. This scatter presumably reflects both variation in the regional heterogeneity of the lung injury between subjects and their response to treatment.
We have presented two typical cases of ARDS which illustrate that these relationships are also generally reflected during acute changes in blood oxygenation and static respiratory system compliance. The fact that the plasma SP-B/SP-A ratio varies with lung function suggests that plasma SP-B is a more sensitive marker of changes in lung permeability than is SP-A. Certainly, numerous studies have shown that the alveolar epithelium recovers rapidly enough after severe injury to alter the sieving properties of the alveolocapillary barrier (6, 39).
While it has been convenient to delineate pulmonary edema as either cardiogenic, usually associated with left heart failure, or noncardiogenic, arising from increases in alveolocapillary permeability, it is becoming increasingly clear that both types share elements in common. The explosive presentation of APE, sometimes described as “flash pulmonary edema,” implies a large, rapid rise in pulmonary capillary pressure which may produce stress failure in the capillaries and increase permeability (6, 7). Therefore, the increase in plasma SP-A and -B in APE patients presumably reflects a similar but less profound increase in alveolocapillary permeability. Indeed, oxidized levels of glutathione are elevated in both APE and ARDS patients (40). Consistent with the studies of Bunnell and Pacht (40), the continuing elevation of plasma SP-A and -B in our patients with APE ∼ 24 h after their entry into the ICU and during resolution of the disease, implies persisting lung injury despite a presumed fall in filtration pressure.
Our finding of a weak relationship between the plasma ratio of SP-B/SP-A and PCP in ARDS suggests that even in classic noncardiogenic edema, permeability is partly related to hydrostatic forces.
In summary, measurement of extravascular lung water is often an insensitive marker of lung damage in patients with ARDS, since the oxygen defect is related more to the ratio of permeability/surface area (41). Moreover, extravascular filtration rate may also be associated with the loss of tissue forces opposing filtration rather than with an increase in permeability (42). Therefore, although increased pulmonary-vascular permeability is considered the hallmark of ARDS, diagnosis for the most part is by clinical inference; noninvasive biochemical markers of lung damage have much to offer. Our studies suggest that plasma SP-B is an acute marker of lung function and alveolocapillary membrane injury. If confirmed, multiple indicators (43), including surfactant proteins, together with current clinical measures of lung function, may find widespread application as noninvasive indicators of damage to the alveolocapillary membrane.
The writers gratefully acknowledge Dr. Malcolm Whiting, Department of Biochemistry and Clinical Pathology, Flinders Medical Centre, for performing the haptoglobin analyses. They thank Dr. Tim Weaver for his helpful suggestions. They further thank Drs. Weaver and Sui Lin for their gifts of the full-length recombinant SP-B and the antibodies 55019 and 96189. They also wish to thank Dr. Jeffrey Whitsett for his generous gift of the antibody 4633.
Supported by Grant 950054 from the National Health and Medical Research Council of Australia.
1. | Dreyfuss D., Saumon G.Role of tidal volume, FRC, and end-inspiratory volume in the development of pulmonary edema following mechanical ventilation. Am. Rev. Respir. Dis148199311941203 |
2. | Kuroki Y., Tsutahara S., Shijubo N., Takahashi H., Shiratori M., Hattori A., Honda Y., Abe S., Akino T.Elevated levels of lung surfactant protein A in sera from patients with idiopathic pulmonary fibrosis and pulmonary alveolar proteinosis. Am. Rev. Respir. Dis1471993723729 |
3. | Chida S., Phelps D. S., Soll R. F., Taeusch H. W.Surfactant proteins and antisurfactant antibodies in sera from infants with respiratory distress syndrome with and without surfactant treatment. Pediatrics8819918489 |
4. | Strayer D. S., Merritt T. A., Lwebuga-Mukasa J., Hallman M.Surfactant-antisurfactant immune complexes in infants with respiratory distress syndrome. Am. J. Pathol1221986353362 |
5. | Honda Y., Kuroki Y., Matsuura E., Nagae H., Takahashi H., Akino T., Abe S.Pulmonary surfactant protein D levels in sera and bronchoalveolar lavage fluids. Am. J. Respir. Crit. Care Med152199518601866 |
6. | Matthay, M. A., G. Nitenberg, and C. Jayr. 1995. The critical role of the alveolar epithelial barrier in acute lung injury. In J. L. Vincent, editor. 1995 Year Book of Intensive Care and Emergency Medicine. Springer-Verlag, Berlin. 28–43. |
7. | Doyle I. R., Nicholas T. E., Bersten A. D.Serum surfactant protein-A (SP-A) levels in patients with acute cardiogenic pulmonary edema and adult respiratory distress syndrome. Am. J. Respir. Crit. Care Med1521995307317 |
8. | Voss T., Eistetter H., Schäfer K. P., Engel J.Macromolecular organization of natural and recombinant lung surfactant protein SP 28-36. Structural homology with the complement factor Clq. J. Mol. Biol2011988219227 |
9. | Johansson J., Curstedt T., Jornvall H.Surfactant protein B: disulfide bridges, structural properties, and kringle similarities. Biochemistry30199169176921 |
10. | Longo M. L., Waring A., Zasadzinski J. A. N.Lipid bilayer surface association of lung surfactant protein SP-B, amphipathic segment detected by flow immunofluorescence. Biophys. J631992760773 |
11. | Weaver T. E., Whitsett J. A.Processing of hydrophobic pulmonary surfactant protein B in rat type II cells. Am. J. Physiol2571989L100L108 |
12. | Doyle I. R., Jones M. E., Orgeig S., Barr H. A., Crockett A. J., McDonald C. F., Nicholas T. E.The ratio of surfactant protein A (SP-A), cholesterol and disaturated phospholipid in human surfactant varies with level of fitness and exercise. Am. J. Respir. Crit. Care Med149199416191627 |
13. | Yogalingam G., Doyle I. R., Power J. H. T.Expression and distribution of surfactant proteins SP-A, SP-B, SP-C and lysozyme in the rat lung following prolonged periods of hyperpnea. Am. J. Physiol141996L320L330 |
14. | O'Reilly M. A., Weaver T. E., Pilot-Matias T. J., Sarin V. K., Gazdar A. F., Whitsett J. A.In vitro translation, post-translational processing and secretion of pulmonary surfactant protein B precursors. Biochim. Biophys. Acta10111989140148 |
15. | Weaver T. E., Sarin V. K., Sawtell N., Hull W. M., Whitsett J. A.Identification of surfactant proteolipid SP-B in human surfactant and fetal lung. J. Appl. Physiol651988982987 |
16. | Lin S., Phillips K. S., Wilder M. R., Weaver T. E.Structural requirements for intracellular transport of pulmonary surfactant protein B (SP-B). Biochim. Biophys. Acta13121996177185 |
17. | Weaver T. E.Single-step purification of recombinant proteins: applications to surfactant protein B. Biotechniques201996804808 |
18. | Anderson L., Anderson N. G.High resolution two-dimensional electrophoresis of human plasma proteins. Proc. Natl. Acad. Sci. U.S.A.74197754215425 |
19. | Voorhout W. F., Veenendaal T., Haagsman H. P., Weaver T. E., Whitsett J. A., Van Golde L. M. G., Geuze H. J.Intracellular processing of pulmonary surfactant protein B in an endosomal/lysosomal compartment. Am. J. Physiol2631992L479L486 |
20. | Masson, P. L. 1986. Molecular parameters of purified human plasma proteins. In F. D. Fasman, editor. CRC Handbook of Biochemistry and Molecular Biology. Proteins, Vol. III. CRC Press, Boca Raton, FL. 242–250. |
21. | Bernard A., Marchandise F. X., Depelchin S., Lauwerys R., Sibille Y.Clara cell protein in serum and bronchoalveolar lavage. Eur. Respir. J5199212311238 |
22. | Nomori H., Horio H., Fuyuno G., Kobayashi R., Morinaga S., Hirabayashi Y.Protein 1 (Clara cell protein) serum levels in healthy subjects and patients with bacterial pneumonia. Am. J. Respir. Crit. Care Med1521995746750 |
23. | Folkesson H. G., Matthay M. A., Weström B. R., Kim K. J., Karlsson B. W., Hastings R. H.Alveolar epithelial clearance of protein. J. Appl. Physiol80199614311445 |
24. | Gregory T., Longmore W., Moxley M., Whitsett J., Reed C., Fowler A., Hudson L., Maunder R., Crim C., Hyers T.Surfactant chemical composition and biophysical activity in acute respiratory distress syndrome. J. Clin. Invest65199119761981 |
25. | LeVine A. M., Lotze A., Stanley S., Stroud C., O'Donnell R., Whitsett J., Pollack M. M.Surfactant content in children with inflammatory lung disease. Crit. Care Med24199610621067 |
26. | Lesur O., Veldhuizen R. A., Whitsett J. A., Hull W. M., Possmayer F., Cantin A., Begin R.Surfactant-associated proteins (SP-A, SP-B) are increased proportionally to alveolar phospholipids in sheep silicosis. Lung17119936374 |
27. | Mason, R. J. 1992. Surfactant secretion. In Robertson, L. M. G. Van Golde, and J. J. Batenburg, editors. Pulmonary Surfactant: From Molecular Biology to Clinical Practice. Elsevier, New York. 306. |
28. | Holter J. F., Weiland J. E., Pacht E. R., Gadek J. E., Davis W. B.Protein permeability in the adult respiratory distress syndrome: loss of size selectivity of the alveolar epithelium. J. Clin. Invest78198615131522 |
29. | Blake, L. H. 1977. Mathematical modelling of steady state fluid and protein exchange in the lungs. In N. C. Staub, editor. Lung Solute Exchange. Marcel Dekker, New York. 99–127. |
30. | DeFouw D. O.Ultrastructural features of alveolar epithelial transport. Am. Rev. Respir. Dis127(Suppl.)1983S9S13 |
31. | DeMello D. E., Heyman S., Phelps D. S., Hamvas A., Nogee L., Cole S., Colten H. R.Ultrastructural of lung in surfactant protein B deficiency. Am. J. Respir. Cell Mol. Biol111994230239 |
32. | Guder W. G.Proteinuria: causes, forms and methods of determination. Clin. Diag. Lab. Med119884955 |
33. | Kabanda A., Jadoul M., Pochet J. M., Lauwerys R., Van Ypersele De Strihou C., Bernard A.Determinants of the serum concentrations of low molecular weight proteins in patients on maintenance hemodialysis. Kidney Int45199416891696 |
34. | Doyle I. R., Morton S., Barr H. A., Crockett A. J., Davidson K. G., Jones M., Jones M. E., Nicholas T. E.Training alters pulmonary surfactant composition (abstract). Am. J. Respir. Crit. Care Med1531996A644 |
35. | Bersten A. D., Doyle I. R., Davidson K. G., Barr H. A., Nicholas T. E.Lung hyperinflation and oxygenation reflect surfactant protein and phospholipid composition in patients with ALI/ ARDS (abstract). Am. J. Respir. Crit. Care Med1531996A587 |
36. | Green K. E., Wright J. R., Wong W. B., Steinberg K. P., Ruzinski J. T., Hudson L. D., Martin T. R.Serial SP-A levels in BAL and serum of patients with ARDS (abstract). Am. J. Respir. Crit. Care Med1531996A587 |
37. | Jobe A., Ikegami M., Jacobs H., Jones S., Conaway D.Permeability of premature lamb lungs to protein and the effect of surfactant on that permeability. J. Appl. Physiol551983169176 |
38. | Suzuki Y., Robertson B., Fujita Y., Grossman G., Kogishi K., Curstedt T.Lung protein leakage in respiratory failure induced by a hybridoma making monoclonal antibody to the hydrophobic surfactant-associated polypeptide SP-B. Int. J. Exp. Pathol731992325333 |
39. | Wiener-Kronish J. P., Broaddus V. C., Albertine K. H., Gropper M. A., Matthay M. A., Staub N. C.Relationship of pleural infusions to increased permeability pulmonary edema in anesthetized sheep. J. Clin. Invest82198814221429 |
40. | Bunnell E., Pacht E. R.Oxidized glutathione is increased in the alveolar fluid of patients with the adult respiratory distress syndrome. Am. Rev. Respir. Dis148199311741178 |
41. | Brigham K. L., Kariman K., Harris T. R., Snapper J. R., Bernard G. R., Young S. L.Correlation of oxygenation with vascular permeability-surface area but not with lung water in humans with acute respiratory failure and pulmonary edema. J. Clin. Invest721983339349 |
42. | Parker J. C., Townsley M. I., Cartledge J. T.Lung edema increases transvascular filtration rates but not filtration coefficient. J. Appl. Physiol66199015531560 |
43. | Pittet J. F., Mackersie R. C., Martin T. R., Matthay M. A.Biological markers of acute lung injury: prognostic and pathogenetic significance. Am. J. Respir. Crit. Care Med155199711871205 |