Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Pulmonary sarcoidosis is a noncaseating granulomatous disorder of unknown etiology. The preferential localization in the lung of activated immune cells, primarily activated oligoclonal CD4+ T cells and macrophages together with the release of various pro-inflammatory cytokines, determines the immune phenomena as well as the development and fate of the sarcoid granulomata (1, 2). Significantly, this can occur in the alveolar, bronchial, or vascular walls. Usually, these granulomata are self-limiting with spontaneous resolution. However, in about 20% of cases the granulomatous inflammation progresses to local or diffuse fibrosis regardless of treatment, with irreversible impairment of lung function (3, 4).

We have previously demonstrated that alveolar macrophages (AM) have a critical influence not only in the initiation of sarcoid inflammation through T cell-mediated immune responses, but also in subsequent repair processes through interactions with other immune effector cells such as fibroblasts (5, 6). Total numbers of AM are raised in bronchoalveolar lavage (BAL) and biopsies of sarcoid patients in addition to phenotypic and functional changes in AM that correlate with clinical status (5, 7). A variety of mediators capable of modulating fibroblast function are released from activated AM, including cytokines (8) and growth factors such as insulin-like growth factor (IGF)-1 (9). IGF-1 is a profibrogenic mediator acting as a potent mitogen and stimulator of collagen synthesis in fibroblasts (10, 11). The AM form of IGF-1 is considerably larger (20-26 kD) than the serum form (7.6 kD) but exhibits the same fibrogenicity (9). As well as AM, other cell types demonstrate expression of IGF-1 transcripts and release of immunoreactive IGF-1 (12). Increased expression and release of IGF-1 has previously been shown in patients with idiopathic pulmonary fibrosis (15) and systemic sclerosis (16).

Local bioavailability of IGF-1 in the lung is, however, influenced by a system of at least six specific high-affinity IGF-binding proteins (IGFBP) 1-6 that act as extracellular reservoirs for maintaining IGF-1 levels (17). IGFBP-3 is the major circulating form (18), which forms a large (150-kD) complex with IGF-1 and an acid-labile subunit, greatly prolonging IGF-1 half-life (19). The functional roles of IGFBP are relatively poorly understood, but have been extensively reviewed recently (17, 20, 21). Their actions are generally inhibitory through interference with IGF-1 binding to its receptor, but can also be stimulatory. This involves association of IGFBP with the cell surface or extracellular matrix, lowering affinity of the IGFBP for IGF-1. In addition, they appear to possess IGF-1-independent growth inhibitory activities, which may be mediated through cell-surface association, perhaps via an IGFBP receptor (22, 23). Post-translational modification of IGFBP also occurs, with phosphorylation increasing and proteolysis decreasing IGFBP affinities for IGF-1, thus inhibiting or potentiating the bioactivity of IGF-1 (24).

Pulmonary sarcoidosis serves as an excellent clinical model for studying not only the initiation of inflammatory response but also events involved in its resolution or progression to fibrosis. However, it remains unclear what pathogenic mechanisms are involved in the lung repair processes, which in some sarcoid patients are overridden, leading to irreversible fibrogenesis. In the present study we have investigated the contribution of IGF-1, and the modulation of its bioactivity by local IGFBP, to the fibrogenesis characteristic of sarcoidosis. For this purpose we have compared the fibroblast mitogenic potential of recovered BAL fluid (BALF) from normal subjects and sarcoid patients (clearly identified as having parenchymal involvement) and assessed the contribution of IGF-1 to the mitogenesis. We have also established an IGFBP profile for normal and sarcoid BALF, and quantified differences in expression of IGFBP-3. IGFBP-3 was selected for its unique spectrum of characterized functions, namely: major serum carrier, cell surface association, IGF potentiation, IGF inhibition, and direct inhibitor of proliferation (17). IGF-1 and IGFBP-3 gene expression were determined in freshly isolated ex-vivo BAL cells (BALC) to confirm that cells associated with the localized immune response could serve as sources for the production of these mediators.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Subject Group

Eight nonsmoking patients (3 male, 5 female) with clinically active pulmonary sarcoidosis were identified after screening by chest radiograph, high resolution CT scan, pulmonary function testing, and transbronchial biopsy. All had evidence of stage III sarcoidosis with bilateral nodular parenchymal infiltration including areas of fibrosis and honeycombing. Three patients had significant reductions in their total lung capacity and all exhibited decreases in carbon monoxide transfer factor. None of the patients had received any prior steroid or other immunosuppressive therapy. The control group consisted of 10 nonsmoking volunteers (6 male, 4 female) with normal chest radiography and pulmonary function. None had suffered any respiratory symptoms in the month preceding the study. Mean subject age is shown in Table 1. Prior approval for this study was obtained from the North Staffordshire Hospital Research Ethics Committee (Staffordshire, UK). All subjects gave informed written consent.

Bronchoalveolar Lavage

BAL was performed on all subjects using a flexible fiberoptic bronchoscope following intravenous midazolam administered 5 min before the procedure. Topical anesthesia, with 2% lignocaine, was applied to the naso-oropharynx and airway. The right middle lobe was instilled with successive 20-ml aliquots of sterile 0.9% isotonic saline, to a volume of 180 ml. The lavage fluid aliquots were immediately gently aspirated by suction, collected into sterile siliconized glass bottles, and maintained at 4°C.

Sample Processing

BALF was centrifuged at 400 × g at 4°C for 5 min and the cell pellet obtained was washed in RPMI 1640 culture medium (Life Technologies, Paisley, UK). Cell viability was assessed by cell exclusion of trypan blue. The cell concentration in RPMI 1640 was adjusted to 1 × 10⁶ cells · ml¹ and 3 × 10⁵ cells removed for cytospin preparations (Cytospin 2; Shandon, Basingstoke, UK). Cytospins were stained for morphology (May and Grunwald stain; Merck Ltd., Lutterworth, UK) and a differential count was performed. The remaining cell suspension was re-centrifuged as before, and the cell pellet was homogenized using 1 ml of Trizol per 10⁶ cells (Life Technologies) and stored at 80°C.

The cell-free BALF supernatant from above was retained and a 1-ml aliquot removed for measurement of urea concentration using a modified microplate-based colorimetric assay (Sigma Diagnostics, Urea Nitrogen; Sigma Chemical Co., Poole, UK). Differences in BALF urea levels between the groups were used to allow dilution effects resulting from the BAL to be taken into account in the subsequent Western and cell proliferation analyses (27). BALF was concentrated 10× by centrifugal filtration using a membrane with a 3-kD cutoff (Amicon, Stonehouse, UK) and stored in aliquots at 40°C.

Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

Total ribonucleic acid (RNA) was purified from the Trizol homogenates according to manufacturers' protocol and resuspended in ribonuclease (RNase)-free water for spectrophotometric evaluation of quantity and purity. Samples A_260/280 were  1.7. RNA was stored in 75% ethanol at 80°C. For complementary DNA (cDNA) synthesis, 1 µg total RNA was used in a 20-µl reaction and incubated for 1 h at 42°C with: 20 pmol oligo dT primer; 0.5 mM deoxynucleotide triphosphates (dNTPs); 0.5 U RNase inhibitor; 50 mM Tris, pH 8.3; 75 mM KCl; 3 mM MgCl₂; and 200 U murine Maloney leukemia virus reverse transcriptase (Clontech; Cambridge Bioscience, Cambridge, UK). The reaction was stopped by heating at 95°C for 5 min, then chilling on ice. cDNA was diluted in 100 µl water and stored at 80°C. A total of 15 µl was used in a PCR.

Specific PCR primers were designed with computer assistance (Primer, version 0.5; Whitehead Institute for Biomedical Research, Cambridge, MA). Oligonucleotides (Table 2) were synthesized commercially (Cruachem Ltd., Glasgow, UK). PCR reactions consisted of 25 pmol primers, 200 µM dNTPs, 1.5 mM MgCl₂, 5 µl 10× PCR buffer, and 2.5 U Amplitaq Gold DNA polymerase (Perkin-Elmer, Warrington, UK) in 50 µl total. A "hot start" PCR was employed in which cycling conditions were: 95°C enzyme activation for 12 min; then 1 min, 15 s at 94°C; 1 min at 55°C; and 1 min at 72°C for 25 cycles (glyceraldehyde-3-phosphate dehydrogenase [GAPDH]) or 40 cycles (IGF-1, IGFBP-3); then a final extension of 10 min at 72°C. Amplified PCR products were resolved on 1.5% agarose gels, stained with SYBR Green I (Flowgen Instruments, Lichfield, UK) and photographed under ultraviolet transillumination with a Wratten 15 filter (Sigma). Specificity of amplified IGF-1 and IGFBP-3 products was verified by: (1) predicted size, (2) restriction digest, and (3) DNA sequencing. For (2), IGF-1 DNA was restricted at a unique PstI site and IGFBP-3 DNA at a unique SphI site by appropriate enzymes (Boehringer Mannheim, Lewes, UK). For (3), the identified band was carefully excised from the gel and DNA isolated and purified using Sephaglas (Pharmacia Biotech, St. Albans, UK). Fragments were blunt-ended with DNA polymerase I (Promega Corp., Southampton, UK) and ligated into the plasmid pUC18 using a Rapid DNA ligation kit (Boehringer Mannheim). These were subcloned into DH5-competent cells according to supplier recommendations (Life Technologies). Recombinant clones were selected and DNA-purified from culture by spin-column (Promega) for direct plasmid sequencing of template by the Sequenase Quick-Denature protocol (Amersham International, Little Chalfont, UK).

Western Ligand Blotting

Ligand blotting was performed as described by Coulson and associates (28). Briefly, 5-µl BALF samples were electrophoresed on 12.5% sodium dodecyl sulfate polyacrylamide gels overnight. Molecular size standards were run alongside samples and comprised: ovalbumin (43 kD), carbonic anhydrase (30 kD), soybean trypsin inhibitor (20.1 kD), and -lactalbumin (14.4 kD). Resolved proteins were electroblotted onto Hybond-C extra filters (Amersham) at 4°C. Filters were blocked with 10 mM Tris, 150 mM NaCl, 0.2% (vol/vol) Tween 20, 5% (wt/vol) dried milk, pH 7.4, and probed with 0.6 × 10⁶ cpm/ml [¹²⁵I] IGF-1 (Amersham) for 2 h, washed, and probe-bound to IGFBP visualized by autoradiography. For reference, normal human serum (NHS) from 10 healthy adults was pooled and 2.5 µl run in parallel with BALF samples on all gels.

Western Immunoblotting

Immunoblotting was carried out using the method of Cwyfan-Hughes and coworkers (29). In short, ligand blots were blocked as described above and then rotated overnight with 1:1,000 dilution of specific IGFBP-3 antiserum (SCH-2/5; Celtrix Pharmaceuticals, Santa Clara, CA) (30), washed to remove excess primary antibody, and finally incubated with 1:4,000 diluted peroxidase conjugated antirabbit IgG antibody (Sigma). Excess secondary antibody was removed by washing and bound conjugate was visualized by chemiluminescent substrate detection (ECL; Amersham) onto film. IGFBP-3 levels on films were quantified by densitometric analysis (GS-670 imaging densitometer, Molecular Analyst image analysis software; Bio-Rad Laboratories, Hemel Hempstead, UK). IGFBP-3 was expressed as the intensity of the relevant band(s) relative to that obtained for the NHS reference, described above.

Cell Proliferation

Normal human lung fibroblasts (IMR90; European Collection of Cell Cultures, Salisbury, UK) were grown in 96-well microplates at a seeding density of 5,000 cells per well for 3 d, humidified at 37°C and 5% CO₂ in air. The culture medium was Dulbecco's modified Eagle's medium supplemented with 10% (vol/vol) fetal calf serum (FCS) and 200 mM L-glutamine (Sigma). After 3 d, complete medium was replaced with low (0.2% vol/vol) FCS medium and cells were incubated, as before, for 48 h. Following this preincubation, medium in each well was then replaced with BALF supernatant concentrate, serially diluted (one-sixteenth to one-one thousand twenty-fourth) with low FCS medium, 10 replicates per dilution. Control wells contained low FCS medium alone. After a further 72-h incubation, fibroblast proliferation was determined using a modified crystal violet dye staining method (31).

Briefly, cells were fixed for 1 h by replacing medium with 4% (vol/vol) formaldehyde in phosphate-buffered saline. Microplates were washed 3 times by immersion in distilled H₂O and air-dried. Cells were stained for 20 min with 100 µl per well of 0.1% (wt/vol) crystal violet in 200 mM 2-N-morpholinoethane-sulfonic acid buffer, pH 6.0, with shaking. Plates were washed, as before, then destained 30 min with 100 µl per well of 10% (vol/vol) acetic acid, with shaking. Plates were then read at 570 nm (Dynatech MR5000; Dynatech Laboratories, Billingshurst, UK). Data were expressed as a percent of change in absorbance from normal or sarcoid groups compared with medium-only controls.

Antibody Blocking

To investigate the specific contribution of IGF-1 to BALF-mediated fibroblast proliferation, neutralizing antibodies were utilized. Fibroblasts were seeded and preincubated as described above. Medium from 10 replicates per experiment was replaced with BALF concentrate, diluted to a one-sixteenth final concentration, previously found to be optimal for maximal stimulation, containing 30 µg/ml goat antihuman IGF-1 (R&D Systems Europe Ltd., Abingdon, UK). Recombinant human IGF-1 (R&D Systems) was included at 5 ng/ml in certain experiments for control purposes. After 72 h further incubation, proliferative activity was evaluated as described above.

Statistics

Comparison of IGFBP-3 immunoblot levels in BALF and cellular differences in BAL between normal subjects and sarcoid patients was performed by the Mann-Whitney test with P < 0.05 taken as significant using Minitab statistical software (Minitab Inc., State College, PA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

BAL Analysis

The BAL characteristics of sarcoid and normal subjects are presented in Table 1. As a group, the patients with pulmonary sarcoidosis had significantly higher total cell counts, relatively reduced percentages but increased absolute numbers of AM, and increased absolute numbers and percentages of lymphocytes and cells per milliliter of BALF than did normal subjects. There was no difference in the lavage fluid recovery or in the levels of the urea reference marker between the two groups. Cell viability as assessed by trypan blue exclusion was > 90% in all samples.

BALF IGFBP Profile

Determination of the IGFBP profile in BALF by Western ligand blotting resulted in the appearance of several bands. For normal subjects (Figure 1A), bands were identified at approximately 24 kD corresponding with IGFBP-4, at 29 kD corresponding with IGFBP-1, and at 33 kD corresponding with IGFBP-2. There was no detectable IGFBP-3, which could be clearly seen in the control human serum as a migrating doublet at approximately 45 and 42 kD (Figure 1A, lane 1). A similar pattern of bands was found in the gel for sarcoid subjects (Figure 1B), although three patients (5, 7, and 13) who had a more aggressive disease phenotype with significant physiological dysfunction, as described under subject group, appeared to have raised IGFBP-2 levels.

View larger version (64K): [in this window] [in a new window]   View larger version (82K): [in this window] [in a new window]   Figure 1.   Western ligand blots (A, B) and IGFBP-3 immunoblots (C, D) obtained with BALF from normal subjects and sarcoid patients. Panels A and C: lane 1 is a normal human serum control; lane 2 is blank; lanes 3-10 are BALF samples from normal subjects. Panels B and D: lane 1 is a normal human serum control; lane 2 is blank; lanes 3 and 4 are BALF samples from normal subjects; lanes 5-7 and 10-14 are BALF samples from sarcoid patients (lanes 8 and 9 are nonsarcoid and not included). Positions of migrating IGFBP are indicated on the ligand blots and the positions of molecular size markers are shown.

IGFBP-3 Immunoblotting

In contrast with the ligand blotting experiments described above, immunoblotting revealed IGFBP-3 staining in BALF samples for both normal (Figure 1C) and sarcoid (Figure 1D) groups. However, IGFBP-3 staining was most prominent with proteins migrating with molecular weights of approximately 29 kD and, less pronounced, in some samples at approximately 16 kD. From a comparison of staining intensity, the 16-kD fragment was more prominent in the sarcoid samples, but no quantitation of this difference was attempted. Both the intact IGFBP-3 doublet and the modified forms were also evident by immunostaining in the control human serum (Figure 1C, lane 1). It can be seen that in most of the BALF samples from both normal and sarcoid groups, the unmodified IGFBP-3 is completely absent, although in some of the samples what appears to be residual 46-kD fragment remains. A quantitative analysis based on the 29-kD IGFBP-3 band was performed (Figure 2) and demonstrated a highly significant increase (P < 0.003) in IGFBP-3 levels in sarcoid BALF compared with normal BALF (no significant difference in BALF urea concentration was found between the groups).

View larger version (49K): [in this window] [in a new window]   Figure 2.   IGFBP-3 (29-kD fragment) in BALF from normal subjects (n = 10) and sarcoid patients (n = 8). Quantitative densitometric analysis of the IGFBP-3 immunoblotting data was performed (see MATERIALS AND METHODS) and data are expressed as means ± SEM for each subject group. Significant difference between groups: *P < 0.003.

RT-PCR Gene Expression of IGF-1 and IGFBP-3 in BALC

RT-PCR of BALC messenger RNA (mRNA) isolated from normal and sarcoid subjects was carried out to confirm the presence of local alveolar sources for these molecules. Representative gels are shown in Figure 3. Single amplification products could be detected for both IGF-1 (Figure 3, bottom panel) and IGFBP-3 (Figure 3, top panel) as well as the GAPDH control (Figure 3, middle panel) in all samples examined. The PCR products in each case correspond with the predicted sizes of 564 (IGF-1), 723 (IGFBP-3), and 983 bp (GAPDH). The identities of IGF-1 and IGFBP-3 products were confirmed by sequence analysis and restriction analysis (not shown). Restriction digest of the IGF-1 product with PstI produced bands of 262 and 302 bp; restriction of purified IGFBP-3 product with SphI gave bands of 430 and 293 bp. These correspond to the known positions of the sites within the amplified region and therefore confirm product identities. Quantitative analysis of gene expression was not performed in this study.

View larger version (42K): [in this window] [in a new window]   Figure 3.   Expression of IGF-1 (bottom panel), IGFBP-3 (top panel), and GAPDH (middle panel) genes in BALC from normal subjects (N) and sarcoid patients (S) detected after RT-PCR (see MATERIALS AND METHODS) on SYBR Green I-stained agarose gels. Representative samples are shown. Sequence-predicted PCR product sizes for each gene are indicated by an arrow. A 100-bp DNA ladder was used as a size reference marker. Gel origins are to the bottom of each panel.

BALF-induced Cell Proliferation

Having shown that sarcoid BALF contains increased levels of partially degraded IGFBP-3, it was necessary to carry out additional experiments to reveal whether there was a concomitant effect on the bioavailability of IGF-1 in BALF. Therefore, experiments were performed to establish the proliferative effect of the BALF on normal human fibroblasts. This was carried out, first, to determine whether there was enhanced cell proliferation in the presence of sarcoid BALF; and second, using IGF-1-specific antibody to investigate whether increased free IGF-1 could account for changes in proliferation.

The effect of BALF from sarcoid patients on fibroblast proliferation compared with that obtained from normal subjects is shown in Figure 4. Sarcoid BALF induced proliferation at all dilutions and was significantly higher (P < 0.0005 from one-sixteenth to one-two hundred fifty-sixth; P < 0.005 from one-five hundred twelfth to one-one thousand twenty-fourth) when compared with the modest proliferative effect observed with the one-sixteenth to one-one hundred twenty-eighth dilutions from the normal group and in comparison with the medium control. Maximal proliferation was achieved at one-sixteenth dilutions for both groups, but sarcoid BALF stimulation was > 60% above medium control and more than 3-fold the level observed for the normal group.

View larger version (28K): [in this window] [in a new window]   Figure 4.   BALF-induced fibroblast proliferation from 10 normal subjects and 8 sarcoid patients. Data are expressed as means ± SEM. Significant difference between sarcoid and control values at each dilution: *P < 0.005; **P < 0.0005.

The effect of neutralizing antibodies (Ab) against IGF-1 on the fibroblast proliferative activity of sarcoid and normal BALF concentrates (one-sixteenth dilution) is shown in Figure 5. Antibodies were shown to be capable of completely blocking IGF-1-induced fibroblast proliferation (P < 0.0002) at the concentration employed. Antibodies alone were shown to have a minimal (5%) effect on cell proliferation. Sarcoid BALF-induced activity was reduced by approximately 40% (P < 0.0005) by Ab; however, there was no significant reduction in the more modest proliferation demonstrated by normal BALF in the presence of anti-IGF-1 Ab.

View larger version (33K): [in this window] [in a new window]   Figure 5.   Effect of IGF-1 antibodies on sarcoid and normal BALF-induced fibroblast proliferation. Results from 10 normal subjects and 8 sarcoid patients are expressed as mean values ± SEM relative to baseline. The additions of BALF (at a dilution of one-sixteenth), recombinant human IGF-1, and anti-IGF-1 antibody were made alone or in the combinations indicated.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Fibrosis is recognized as the end stage of a complex process of abnormal repair. Identification of consistent cellular responses to injury leading to development of fibrosis may lead to earlier and more successful therapeutic intervention (32). Progression of many interstitial lung diseases leads ultimately to lung fibrosis, although it is still unclear why, in pulmonary sarcoidosis, fibrosis develops only in a relatively small proportion of patients. Characteristically, the preceding inflammatory response is dominated by an interaction between activated macrophages and lymphocytes. Evidence exists for the production of progression-type fibroblast-stimulating growth factor, IGF-1, by these cell types. Resting peripheral blood lymphocytes have been shown to express IGFBP-2 and IGFBP-3, but after activation IGFBP-5 and IGF-1 are also expressed (12). It is likely, therefore, that expression of IGF-1 and these three IGFBPs would be similarly enhanced during sarcoid inflammation. AM from sarcoid patients spontaneously release an AM-derived growth factor (33), identified as tissue IGF-1, a larger form of the same molecule as serum IGF-1 (11). This comprises the 70-amino-acid mature peptide (exons 3 and 4) with additional N-terminal signal sequence arising from alternative splicing of exons 1 or 2, and a C-terminal E-peptide sequence formed by alternative splicing of exons 4-6. BALC in this study express IGF-1Eb (exons 4-5) transcripts, as shown by RT-PCR. The regions of the IGF molecule that interact with IGFBP are at precise locations within the mature peptide and do not require the presence of prohormone extension sequences for binding. N-terminal-specific extensions lie upstream of the signal peptide and would be lost during processing before any interaction with IGFBP. However, there may be some sort of prior recognition step of prohormone E-domain extension forms that could influence subsequent interactions with IGFBP (34). Increased spontaneous release of IGFBP-3 from AM has been observed in patients with idiopathic pulmonary fibrosis (IPF) (15) but to our knowledge has yet to be described in pulmonary sarcoidosis.

In this paper we have shown by ligand and immunoblotting that intact 46-kD IGFBP-3 is largely absent from normal and sarcoid BALF. However, immunoblotting reveals that both contain IGFBP-3 fragments of 29 and 16 kD with marked increases in expression of these fragments, especially the 29-kD form (see Figure 2), in BALF from the sarcoid patients. The 29-kD IGFBP-3 fragment comigrates with IGFBP-1, and the low signal seen in the ligand blot corresponding to this molecular weight may represent residual IGF-1 binding by the IGFBP-3 fragment rather than by IGFBP-1. The low levels of IGFBP-3 detected by ligand blotting, even in sarcoidosis where IGFBP-3 fragment concentrations are increased, therefore corresponds to a reduction in its affinity for IGF-1 following proteolysis to smaller immunoreactive forms in both normal and sarcoid BALF. Limited proteolysis of circulating and local IGFBP by IGFBP-specific enzymes has been described for IGFBP-2 (35), IGFBP-3 (26, 36), IGFBP-4 (38, 42, 43), and IGFBP-5 (43, 44). It appears that proteases act to lower the affinity of IGF-binding to IGFBP, which may facilitate IGF bioavailability (45). Our data are in agreement with these findings and are suggestive of IGFBP-3 protease activity in epithelial lining fluid which may act locally to regulate free IGF-1 concentration. The modified IGFBP-3 detected in BALF was present on immunoblots as 29- and 16-kD fragments. The 29-kD immunoreactive fragment of IGFBP-3 has been observed in a number of sera and body fluids where protease activity has been detected and is likely to be a proteolytic fragment of IGFBP-3 with reduced affinity for IGFs (46). Recently, a 16-kD non-glycosylated proteolytic fragment of IGFBP-3 has been identified from in vitro cultures of IGFBP-3 with plasmin, a calcium-dependent serine protease (39). Indeed, this enzyme has been identified previously as an IGFBP-3 protease in biologic fluids (47). Other well-known IGFBP proteases include prostate-specific antigen (48) and metalloproteases (49, 50), but clearly there are other, as yet undefined, perhaps IGFBP-specific proteases involved. IGFBP-3 protease activity is increased during gestation (37, 38) and in some pathologic conditions (46), and occurs outside of the bloodstream, even in the normal state (51). It is unclear whether there is local production of protease at the inflammatory site or whether serum proteolysis is responsible.

Production of IGFBP-3 and IGF-1 does occur at local sites, as shown by their BALC gene expression in the lung in the present study. It is therefore possible that AM and lymphocytes are the sources of the IGFBP and IGF-1 rather than the cells forming the architecture of the lung. Indeed, we can confirm that AM separated from unfractionated BALC are a source (results not shown) and it is highly likely that lymphocytes are as well, at least in sarcoid patients, given that activated lymphocytes express both IGF-1 and IGFBP-3 (12). In addition, we have found expression of IGFBP-1, IGFBP-2, IGFBP-3, IGFBP-4, IGFBP-6, and IGFBP-7 in unfractionated BALC from normal and sarcoid patients (results not shown). Quantitative analysis of the expression of IGF-1 and IGFBP-3 mRNA between sarcoid and normal subjects has yet to be performed. However, the significance of any differences between gene expression may be diminished by extensive regulation at the post-translational level, by IGFBPs for IGF-1 and IGFBP proteases for IGFBPs, determining their contributions to the fibrogenic response. Similar observations were recently reported for patients with IPF (15). Interestingly, IGFs can also modulate IGFBP levels, enhancing IGFBP-3 by a mechanism independent of de novo synthesis and by interaction with the IGF-1 receptor, involving release of membrane-associated IGFBP-3 (52). Thus, increased local IGF-1 could result in the enhanced IGFBP-3 reported here for the sarcoid patients. The presence of IGFBP-3 protease would serve to regulate IGF-1 bioavailability further through partial proteolysis of IGFBP-3 and the reduced affinity of the degraded (29-kD) fragment for IGF-1. The 16-kD fragment observed here has been reported to have no affinity at all for IGF (39). Further evidence of enhanced IGF-1-mediated fibrogenic activity in sarcoid patients comes from cell-proliferation studies. Our data showed fibroblast mitogenic potential of the sarcoid BALF to be three times that observed for normal subjects. Use of IGF-1-specific Ab showed that approximately 40% of fibrogenic activity in sarcoid BALF could be attributed to an IGF-1-mediated effect. This contrasted with normal BALF, in which anti-IGF-1 caused no significant reduction in proliferation, suggesting that the IGF-1 component here was very small. These data are in agreement with similar results obtained with BALF from patients with systemic sclerosis (16) and pneumonectomized rats (56).

The role of IGFBP (particularly IGFBP-3) in cell proliferation is, however, more complex than preventing IGF binding to its receptor. Under some conditions it has been suggested that they may potentiate IGF-1 action as well through association of IGFBP with the cell surface or extracellular matrix, lowering affinity of the IGFBP for IGF-1 (17). Recent studies suggest that there is a mechanism for IGF-1-independent inhibition of cell proliferation by IGFBP-3, involving apoptosis, putatively via IGFBP-3 receptors. Crucially, this relies on IGFBP-3 interaction with the cell surface, since complexed IGFBP:IGF-1 does not inhibit growth. Further, the finding of nuclear localization of IGFBP-3 may mean that it is involved in transcriptional control of growth-regulating genes (57). In this study it is possible that increased local production of IGF-1 invokes an IGFBP-3 synthesis response in, and/or release from, cells. This defensive action would lead to sequestration of IGF-1 and limited bioavailability. However, the observed increase in IGFBP-3 in sarcoid patients is met by partial proteolysis and defeat of the cells' anti-proliferative machinery. This scenario simply does not arise in the normal subjects, although they too have IGFBP-protease activity, because this process represents a response to increased free IGF-1 and not a cause of it. Interestingly, in support, purified 30-kD proteolytic fragment has been shown to potentiate IGF-1 action, suggesting that the truncated fragment may have intrinsic mitogenic activity. This was in contrast to inhibition shown by native IGFBP-3 (58). We believe that differential regulation of IGF-1 by associated IGFBPs and proteases may well contribute toward enhanced fibrogenesis and matrix deposition observed in individuals with pulmonary sarcoidosis and could thus be also implicated in the pathogenesis of other interstitial lung diseases. We hypothesize that fibrogenic mediators (IGF-1) are resident in lung and are increased following lung injury as part of the normal wound-repair process, but in fibrogenesis the usual regulatory brakes (putatively IGFBP) fail to halt their expression. Future efforts will be directed toward developing a clearer understanding of interactions between the IGF axis and other pro-inflammatory mediators and fibrogenic growth factors in order to evaluate differences between sarcoid patients with a more benign, resolving, or therapy-responsive disease and those with an aggressive, unresponsive form leading to development of fibrosis. There are many potential modulators of this cascade of events, including growth hormone (GH). Accumulating evidence suggests GH-releasing hormone and GH, as well as IGF-1, are produced from inflammatory cells and have paracrine and autocrine immunomodulatory functions that suggest their involvement in inflammatory pathologies (59). We believe GH is a prime candidate for study because it is a common regulator of IGF-1, IGFBP-3 (60), and perhaps IGFBP-protease. Multiple mechanisms of action and involvement in the regulatory processes balancing normal lung repair against excessive lung remodeling are clearly the features that need to be exploited for the development of new preventive and therapeutic interventions.