| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
*
Department of Cell Biology and Anatomy and Division of Hematology/Oncology,
Department of Medicine, New York Medical College, Valhalla, NY 10514
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
How does IL-6 exist in blood? How is the transport of IL-6 in blood regulated? Are there clinical situations characterized by abnormalities of IL-6 transport in blood? Despite previous reports suggesting that IL-6 may exist in human blood solely as a monomeric 25- to 30-kDa protein with B9 hybridoma cell proliferation activity (3, 4), several lines of investigation have pointed to a more complex regulation of IL-6 transport in human blood.
Plasma/sera derived from normal volunteers analyzed by Sephadex G-200 gel filtration chromatography contained low levels of IL-6 (0.5–5 ng/ml) in the form of complexes of 100 to 150 and 400 to 500 kDa (5). These IL-6 complexes in normal plasma/serum were reactive in the 4IL6/5IL6 ELISA developed by us, but were nonreactive in an otherwise highly sensitive IL-6 ELISA (Ig61/5IL6 ELISA; sensitivity using Escherichia coli-derived IL-6, 1–10 pg/ml) and were devoid of B9 hybridoma cell proliferation activity (5). Higher levels of plasma/serum IL-6 (10–100 ng/ml) in the form of 100- to 150-kDa and 400- to 500-kDa IL-6 complexes with 4IL6/5IL6 ELISA reactivity, but with no B9 bioactivity, were observed in patients with epidermolysis bullosa and psoriasis (5). In contrast, serum from a patient with acute infection contained, in addition to the 100- to 150-kDa and 400- to 500-kDa IL-6 complexes, 25- to 30-kDa B9-bioactive and Ig61/5IL6 ELISA-reactive IL-6, although the B9 bioactivity in this serum sample (estimated at 300 U/ml) was approximately 1000-fold lower than would have been expected based upon the concentration of IL-6 in this serum sample (5 µg/ml) as estimated by immunoaffinity purification and direct amino acid sequencing (5).
The influence of anti-IL-6 Abs in regulating IL-6 transport in blood became evident through experimental studies in mice, baboon, and man in which the administration of anti-IL-6 mAb led to the appearance of long-lived IL-6/anti-IL-6 IgG complexes of approximately 150 to 200 kDa with, typically, a paradoxical enhancement of in vivo IL-6 biologic activity (6, 7). The detection of anti-IL-6 autoantibodies in human sera (8, 9) and of IL-6/sIL-6R3 complexes in sera of patients with juvenile rheumatoid arthritis (10) further pointed to the complexity of IL-6 transport in blood. Additionally, sIL-6R was found in association with IL-6 in long-lived, 200-kDa IL-6 complexes observed in sera from melanoma patients actively immunized with an anti-idiotypic mAb (MK2–23; this mAb bears the internal image of a high m.w. melanoma-specific Ag) (11).
An approach to understanding the complexity of IL-6 transport in human blood and to begin to elucidate its biologic consequences would be the identification of specific clinical settings consistently characterized by specific distinct patterns of IL-6 transport. The identification and investigation of such clinical situations would permit 1) the biochemical and immunologic evaluation of various subsets of circulating IL-6, 2) studies of the regulation of IL-6 transport in blood, and 3) an evaluation of the biologic and immunologic consequences of such abnormalities. We have previously referred to circulating IL-6 present in high molecular mass complexes as "chaperoned" IL-6 to draw attention to this additional level of regulation of IL-6 function in vivo, that of regulated cytokine transport in the intravascular compartment (5, 6, 11).
Prompted by our prior observation that sera from patients subjected to immunization with an active anti-idiotypic mAb immunization regimen (mAb-KLH+BCG) (11) contained high levels of 200-kDa IL-6 and the availability of banked sera from melanoma patients who had been subjected to different active immunization regimens for their cancer, we investigated IL-6 transport and bioavailability in the blood of patients administered different anti-cancer immunizations. Remarkably, distinct classes of chaperoned circulating IL-6 of 30, 200, and 450 kDa, each with distinctive immunologic and biologic availability, were systematically observed in different groups of patients. The present data suggest the existence of specific mechanisms involving distinct binding proteins that regulate IL-6 transport and function in vivo.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Serum was prepared from blood (5–10 ml) that had been collected from adult ambulatory melanoma patients enrolled in various active immunotherapy protocols in late 1995 and early 1996 (n = 57) in the Division of Hematology-Oncology, Department of Medicine, New York Medical College (Valhalla, NY). Upon collection, the blood samples had been kept at 4°C until transported to the laboratory for serum preparation (within 1–20 h of collection). Group I patients (n = 11) did not receive any form of immunotherapy. Group II patients (n = 20) were treated with the anti-Id mAb MK2–23 coupled to KLH together with BCG as adjuvant according to the vaccination protocol described previously (11, 12, 13) (mAb-KLH+BCG). For patients in this group, the last booster of the mAb-KLH vaccine had been administered in or before 1993 (i.e., >2 yr earlier). Group III patients (n = 21) received an autologous anti-cancer Ag preparation (AAAP) (14). Briefly, this procedure involved s.c. vaccination with a zinc chloride-modified tumor cell membrane fraction derived from homologous cancer tissue previously removed during surgery (14). AAAP was administered twice weekly (1 ml/injection) for a period of 9 wk. Group IV patients (n = 5) received both AAAP and mAb-KLH+BCG, but at different times. While three patients were first treated with the mAb-KLH+BCG and then with the AAAP regimen, two patients received AAAP first and then mAb-KLH+BCG. All immunotherapy protocols were approved by the institutional committee for protection of human subjects, which functioned as the institutional review board.
Sephadex G-200 gel filtration
Serum samples (1 ml) were fractionated through a Sephadex G-200
column (2.5 x 60 cm; bed volume (Vo) = 164 ml;
excluded volume (Ve) = 55 ml; included volume
(Vi) = Ve - Vo) as described
previously (5, 6, 11). Eluate fractions of 3 ml were collected after
the first Vo, and the elution was continued for an
additional two Vos (
40 fractions). All eluted fractions
were calibrated for molecular size by use of the following marker
proteins: ribonuclease, 14 kDa; chymotrypsinogen A, 25 kDa; OVA, 43
kDa; aldolase, 158 kDa; catalase, 232 kDa; and ferritin, 440 kDa.
Sephadex G-200 and the calibrating proteins were purchased from
Pharmacia LKB (Piscataway, NJ).
Immunoaffinity chromatography
IL-6, sIL-6R, and their associated proteins present in serum samples were purified by immunoaffinity chromatography using sets of three murine mAb affinity columns (Hydrazide Avidgel, Unisyn Technologies, Inc., Hopkinton, MA) prepared according to manufacturer’s instructions (5, 6, 11). Briefly, immunoaffinity resins were prepared using the anti-IL-6 mAb 5IL6-H17, the anti-IL-6R mAb MT-18, and an irrelevant mAb (0.2 mg/1 ml resin) and packed in 1- to 2-cm columns in Pasteur pipettes. Separate aliquots (0.25 ml) of each serum sample were passed through each of the three columns, the columns were washed extensively with 150 mM NaCl and 10 mM Tris-Cl, pH 7.5. The adsorbed proteins were eluted upon addition of 0.1 M glycine buffer, pH 2.4, and the eluted fractions (0.5 ml each; E1 through E4) were immediately neutralized with 50 µl of 1 M Tris-Cl, pH 8.0. Control experiments verified that there was little or no leaching of the respective murine mAb from these immunoaffinity columns (data not shown).
ELISAs for IL-6
Various ELISAs for human IL-6 were conducted as described previously (5, 6, 11). Two of the ELISAs have been developed by us using the murine mAb pairs 1) 4IL6-H11 for capture and biotinylated 5IL6-H17 for readout (the 4IL6/5IL6 ELISA), and 2) 7IL6-H12 for capture and biotinylated 5IL6-H17 for readout (the 7IL6/5IL6 ELISA) (5, 6, 11). The World Health Organization reference standard 89/548 for human IL-6 was used to calibrate these ELISAs. Additionally, we used three commercial ELISAs for human IL-6 purchased from R&D Systems, Inc. (Minneapolis, MN), Genzyme Corp. (Cambridge, MA), and Endogen, Inc. (Cambridge, MA). These commercial ELISAs were calibrated using the IL-6 standards provided by the respective manufacturers in their ELISA kits. All ELISAs performed on unfractionated serum samples were run in triplicate, whereas all ELISAs on Sephadex G-200 eluate fractions were conducted in duplicate.
Bioassays for IL-6
Two bioassays for IL-6 were used. These were 1) the B9 hybridoma
cell proliferation assay and 2) the Hep3B hepatocyte stimulation assay,
as described previously (5, 6, 11, 15, 16, 17). The procedure for SDS-PAGE
analyses of [35S]methionine-labeled
1-antichymotrypsin secreted by Hep3B cells in response
to IL-6 addition has been described previously (16, 17). All bioassay
data were calibrated in terms of the World Health Organization
reference standard 89/548 for human IL-6.
ELISA for sIL-6R
An ELISA for human sIL-6R was developed by us using a rabbit polyclonal anti-sIL-6R Ab prepared against baculovirus-expressed sIL-6R as the capture Ab and the murine mAb MT18 as the biotinylated readout Ab (11). The sIL-6R ELISA was calibrated using a laboratory reference standard for baculovirus-expressed recombinant sIL-6R (11).
ELISAs for total IgG, anti-IL-6 Ab, and anti-IL-6R Ab
Total human IgG in samples was assayed using an ELISA procedure described by Kenney et al. (18). A modification of the procedure reported by May et al. (5) was used to detect Ig with the ability to bind IL-6 or sIL-6R immobilized on ELISA plates. Recombinant baculovirus-expressed human IL-6 (15 µg/ml) or human sIL-6R (10 µg/ml) was bound to 96-well ELISA plates (100 µl/well). After blocking with 5% nonfat milk (Carnation, Los Angeles, CA) in PBS for 2 h, aliquots (50 µl) of test samples that might contain IL-6- or sIL-6R-binding Ig were added to duplicate wells. Murine serum protein-adsorbed biotinylated rabbit anti-human IgG(H+L) (Southern Biotechnology Associates, Inc., Birmingham, AL) was used as the readout Ab. Additionally, biotinylated murine anti-human IgG1 and IgG2 were used to determine the isotype of the ligand-binding human Ig. Controls for IgG specificity included parallel binding assays using biotinylated anti-human IgA and biotinylated anti-human IgA Abs as readout Abs.
Statistical evaluation
All statistical analyses were performed using the TRUE EPISTAT software package (Epistat Services, Richardson, TX).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Blood was obtained from ambulatory melanoma patients
(n = 57) subjected to different active
anti-cancer immunization regimens. For evaluation of serum IL-6
levels, these patients were subdivided into four groups as follows:
group I (n = 11), patients who did not receive
any immunostimulating therapy; group II (n =
20), patients who had received active immunization with anti-Id mAb
MK2–23 coupled to KLH as carrier together with BCG as adjuvant before
or in 1993 (mAb-KLH+BCG); group III (n = 21),
patients who had been immunized using AAAP; and group IV
(n = 5), patients who had received both AAAP and
mAb-KLH+BCG. Serum IL-6 levels were assayed using the 7IL6/5IL6 ELISA
developed by us earlier (6, 11). In this ELISA, "normal" serum IL-6
values range from 10 to 100 pg/ml (11) (our unpublished observations).
Figure 1
A shows that dramatic
elevations of serum IL-6 levels persisted in many of the mAb-KLH+BCG
patients >2 yr after cessation of the active immunization (group II)
(11). Marked elevations of serum IL-6 levels were also observed in many
of the AAAP patients (group III) and in those given both mAb-KLH+BCG
and AAAP (group IV). There appeared to be a bimodal distribution of
serum IL-6 levels in group II and III patients, in that a subset of
patients within each group had very high IL-6 levels. The data in
Figure 1B show that there was little difference in serum
sIL-6R levels among these four groups of sera. In terms of
stoichiometry, the concentrations of IL-6 in group II, III, and IV sera
were far in excess of those of sIL-6R (compare data in Fig. 1, A and B; also see Fig. 5), suggesting
abnormalities of IL-6 transport in these patients in addition to simply
the formation of IL-6/sIL-6R complexes.
|
|
To investigate the biochemical nature of IL-6 in sera of patients
in groups II, III, and IV, four serum samples from group II and two
samples each from groups III and IV, which contained the highest IL-6
amounts (as in Fig. 1
A), were size fractionated
through a Sephadex G-200 gel filtration column. The immunologic and
biologic properties of IL-6 in the gel filtration eluate fractions were
probed using five different ELISAs for IL-6 and two different bioassays
for IL-6. Additionally, the gel filtration characteristics of sIL-6R
present in these serum samples were probed using an ELISA for sIL-6R.
One representative analysis of IL-6 in serum from each of groups II,
III, and IV is illustrated in Figures 2-4.
|
Figure 2 summarizes the characteristics of circulating IL-6 in a
patient who received mAb-KLH+BCG; similar data were obtained in
analyses of the additional three sera (data not shown). Figure 2A confirms our previous observation (11) that IL-6
circulated in such patients mainly as a 200-kDa complex. IL-6 in this
200-kDa complex was preferentially reactive in the 4IL6/5IL6 ELISA and
less reactive in the 7IL6/5IL6 ELISA (Fig. 2A),
essentially unreactive in the R&D and Genzyme ELISAs for IL-6 (Fig. 2B), and had minimal reactivity in the Endogen ELISA
for IL-6 (Fig. 2B). Figure 2C shows that
the 200-kDa IL-6 was devoid of any detectable biologic activity
in both the B9 hybridoma proliferation and the Hep3B hepatocyte
stimulation bioassays for IL-6.
Characterization of serum IL-6 in patients receiving AAAP (group III)
IL-6 in sera of patients receiving AAAP was 30 and 450 kDa, was
preferentially reactive in the 7IL6/5IL6 ELISA, and was poorly reactive
in the 4IL6/5IL6 ELISA (Fig. 3A). For comparison, it
should be noted that Escherichia coli-derived recombinant
human IL-6 of 21 kDa chromatographs at approximately 13.5 kDa through a
Sephadex G-200 column (20) (data not shown).
|
C). The
three commercial ELISAs for IL-6 had a reduced ability to detect the
30- and 450-kDa IL-6 complexes (Fig. 3B). Characterization of serum IL-6 in patients administered both mAb-KLH+BCG and AAAP (group IV)
The concept of regulated bioavailability of IL-6 from within
distinct and different circulating complexes was emphasized by analyses
of serum IL-6 from two patients who had received both forms of active
anti-melanoma immunotherapy (group IV). The IL-6 chromatography
profiles in Figure 4 are clearly a
composite of Figures 2 and 3. In group IV patients, Figure 4 reveals
the presence in serum of IL-6 complexes of 30, 200, and 450 kDa, with
each complex possessing its characteristic ELISA reactivity and
bioactivity pattern. The 30- and 450-kDa IL-6 complexes were reactive
in the 7IL6/5IL6 and R&D ELISAs, whereas the 200-kDa IL-6 was
preferentially reactive in the 4IL6/5IL6 ELISA, reactive in the
7IL5/5IL6 ELISA, but nonreactive in the R&D ELISA (Fig. 4, A
and B, and inset in A). While the 30- and
450-kDa IL-6 complexes displayed B9 and Hep3B biologic activity, the
200-kDa IL-6 was devoid of activity in these bioassays (Fig. 4C). The composite IL-6 gel filtration profile
depicted in Figure 4 provides clear evidence for the existence of
distinct and different forms of IL-6 transport in blood and the complex
regulation of the bioavailability of this cytokine.
|
Free monomeric sIL-6R is approximately 50 to 60 kDa (reviewed in
Refs. 2, 10, and 19). We evaluated whether sIL-6R in group II, III, and
IV sera behaved as a monomer by gel filtration analyses. The data in
Figure 5 show that in patients who most
recently received AAAP, the bulk of sIL-6R was 150 kDa with,
additionally, a distinct peak at approximately 450 kDa. Thus, these
sera contained little free sIL-6R (although a shoulder at 50–60 kDa is
evident in sample 511), consistent with the demonstration in Figure 1, A and B, that from a stoichiometric standpoint
there was a vast excess of IL-6 in these sera that would have bound all
the sIL-6R present. The molecular mass of the major sIL-6R peak (150
kDa) is consistent with an expected trimeric complex of the form
[IL-6][sIL-6R]2 (19). The 450-kDa sIL-6R peak in AAAP
serum is coincident with the 450-kDa IL-6 complex (see Figs. 3 and 4)
and is consistent with the inclusion of sIL-6R in these 450-kDa IL-6
complexes. It has been previously shown that recombinant preparations
of IL-6, sIL-6R, and sgp130 can be artificially reconstituted into a
440-kDa hexameric complex,
[IL-6]2[sIL-6R]2[sgp130]2
(19); however, IL-6 bioactivity, as assayed in B9 and HepG2 hepatoma
cells, is inhibited in such artificially reconstituted complexes (Ref.
20 and citations therein). The observation that the endogenous
serum-derived 450-kDa complex displayed B9 and Hep3B bioactivity
commensurate with 7IL6/5IL6 ELISA reactivity compared with that of the
30-kDa IL-6 (Figs. 3 and 4) suggests that the serum-derived 440-kDa
complex may be structurally different from the artificially
reconstituted
[IL-6]2[sIL-6R]2[sgp130]2
complex.
Sera from patients receiving mAb-KLH+BCG alone (samples 524 and 582)
also contained little free monomeric sIL-6R (Fig. 5). Compared with the
sharp 150-kDa sIL-6R peak in sera from AAAP patients, there was a
reproducible shift in sIL-6R gel filtration properties toward a complex
of 200 kDa. The data in Figure 2, showing a 200-kDa IL-6 complex, and
the demonstration of anti-IL-6R-binding and anti-IL-6-binding
autoantibodies in such sera (see below) suggest that this shifted
200-kDa sIL-6R peak (Fig. 5 in samples 582 and 524) is consistent with
Ab-containing complexes of the form [IL-6][sIL-6R][Ab] and
[Ab][IL-6][sIL-6].
It is noteworthy in Figure 5 that the sIL-6R-containing complexes in
patients receiving mAb-KLH+BCG were shifted to a size >440 kDa,
suggesting that this protocol generated >440-kDa biologically inactive
IL-6 complexes different in composition from the biologicly active
440-kDa IL-6 complexes seen in AAAP patients (also compare the
>440-kDa peak in Fig. 2 with the 440-kDa peak in Fig. 3).
Ligand-occupied anti-IL-6- and anti-sIL-6R-binding autoantibodies in group II patients
The lack of IL-6 biologic activity in ex vivo assays and the
200-kDa molecular mass estimate for IL-6 by gel filtration analysis
were reminiscent of the 150- to 200-kDa IL-6/anti-IL-6 Ab complexes
observed in the circulation of mice, baboons, and man administered
anti-IL-6 mAb (6, 7). We first used the procedure of Bendtzen et
al. (8, 9) to detect anti-IL-6 autoantibodies in these sera, using
50-µl aliquots of sera to immunoprecipitate
125I-labeled rIL-6. Our failure to detect
anti-IL-6-binding autoantibodies in group II sera using this assay
(data not shown) suggested that if such binding Abs were present, that
the ligand binding site might be fully occupied by IL-6 (consistent
with the 200-kDa gel filtration property of the IL-6 ligand; Fig. 2)
with little free anti-IL-6-Ig available for additional detection
using the [125I]rIL-6 immunoprecipitation assay.
Therefore, an alternative approach to detect ligand-occupied
autoantibodies to IL-6 or sIL-6R in these sera was developed (Fig. 6).
|
)
containing the following immobilized murine mAb: 1) anti-IL-6 mAb
5IL6-H17 (this mAb is reactive with 200-kDa IL-6; see Fig. 2), 2) a
control isotype-matched mouse mAb (IgG1), or 3) the anti-sIL-6R mAb
MT18. After extensive washing, column-bound IL-6 or sIL-R together with
their associated proteins (including, presumably, ligand-bound human
Igs) were eluted using acidic glycine buffer (pH 2.4) followed by
immediate dilution and neutralization by 1 M Tris (pH 8.0). It was
expected that the acid elution step would not only release the IL-6 and
sIL-6R ligand from the respective mAb comprising the immunoaffinity
matrix (as shown in Refs. 5, 6, and 11), but would also release any
ligand that occupied the binding site on anti-IL-6- or
anti-sIL-6R-binding autoantibodies indirectly retained on the
columns. These unoccupied human Igs in the diluted eluate would
then be available for detection in ELISA plates containing immobilized
IL-6 or sIL-6R. We have previously reported characterization of the
IL-6 and sIL-6R that are purified using this procedure from sera of
patients administered mAb-KLH+BCG (see Fig. 6 in 11 .
Human IgG bound to and eluted from the anti-IL-6-mAb and
anti-IL-6R mAb columns using JS serum (Fig. 6A).
JS serum also contained human IgG that bound to the control irrelevant
murine mAb column (Fig. 6A), indicative of the
development of a human anti-mouse Ig Ab (HAMA) in this patient who
had been immunized with the murine mAb MK2–23 (12, 13).
The quantity of human anti-IL-6-binding or anti-IL-6R binding
IgG in the respective column eluates was estimated by allowing these
eluate fractions to bind to either IL-6 or sIL-6R Ags immobilized on an
ELISA plate followed by detection of the bound IgG using biotinylated
human-specific anti-IgG reagents. IL-6-binding and sIL-6R-binding
IgG were observed in eluates from JS serum (Fig. 6, B and
C; expressed as arbitrary absorbance units for each panel).
In contrast, only low levels of sIL-6R-binding IgG were detected in RC
serum (Fig. 6, E and F). Data similar to
those for JS in Figure 6 were obtained in analyses of four additional
sera from the mAb-KLH+BCG group II, and those similar to data for RC
were obtained in analyses of four additional sera from the control
group I (not shown). Additionally, the predominant isotype of the IL-6-
and sIL-6R-binding Ig in sera from group II patients was determined to
be IgG1 (not shown). Taken together, data such as those presented in
Figure 6 are consistent with a model in which the 200-kDa complexes
comprise populations of binary ligand-occupied anti-IL-6 or
anti-sIL-6R IgG as well as of ternary complexes of the type
IL-6/sIL-6R/Ab and Ab/IL-6/sIL-6R (consistent, respectively, with
purification of anti-sIL-6R-binding IgG off the anti-IL-6 mAb
column 1 as in C and of anti-IL-6-binding IgG off the
anti-sIL-6R mAb column 3 as in B). The
Ab-containing binary complexes would not be expected to be resolved
from the ternary complexes by Sephadex G-200 chromatography as
used in this study.
A practical implication: severe discrepancies in serum IL-6 data derived using different ELISAs for IL-6
If the different ELISA reactivities of IL-6 complexes separated by
Sephadex G-200 gel filtration as depicted in Figures 2 to 4 hold true
for IL-6 present in the respective unfractionated serum samples, then
one can predict that there will arise severe discrepancies in
descriptions of serum IL-6 levels using different ELISAs. Figure 7A illustrates examples in
which IL-6 levels in sera assayed using the 7IL6/5IL6 ELISA were
compared with those estimated using the commercial R&D ELISA, and
Figure 7B presents a similar comparison between data
obtained using the 4IL6/5IL6 ELISA and the commercial R&D ELISA. It is
clear that the commercial R&D ELISA grossly underestimates the level of
serum IL-6 in many, if not all, of these patients. In additional
comparisons, the Genzyme and the Endogen ELISAs also provided severe
underestimates of serum IL-6 levels (data not shown).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The identification of several distinct abnormalities in IL-6 transport in blood emphasizes the complexity of the intravascular pool of this cytokine. That the particular active immunization protocol used determines which of these complexes predominates in blood points to the independent regulation of each transport modality. The present observations find a parallel in the literature concerning insulin-like growth factor binding protein (IGFBP); the intravascular bioavailability of insulin-like growth factor (IGF-I) is negatively or positively regulated by six different binding proteins (IGFBP-1 through -6), each of which is independently expressed in different tissues (21, 22, 23, 24). The bioavailability of IGF-I from intravascular complexes with respective binding proteins is, in turn, regulated by specific proteases that cleave the binding proteins, reducing the latter’s affinity for the ligand (21, 22, 23, 24). As an example, IGFBP-1 has an inhibitory effect on IGF-I function because it sequesters IGF-I in the vascular compartment; release of IGF-I from IGFBP-1 requires a specific IGFBP-1 protease. In contrast, binding of IGF-I to IGFBP-3 or -4 enhances IGF-I function because these binding proteins enhance the transit of IGF-I out of the vascular compartment, increasing the delivery of IGF-I to target tissues. Proteases specific to each of these binding proteins are then involved in releasing IGF-I to the target tissues. Characteristic abnormalities in the levels of circulating IGFBP and of IGF-I transport occur in specific disease states, such as an increase in circulating IGFBP-1 in diabetes, a decrease in circulating IGFBP-1 in hyperinsulinemia of obese menopausal women, a decrease in IGFBP-2 in growth hormone deficiency, and an increase in IGFBP-3 in acromegaly (reviewed in Refs. 23 and 25).
It has become apparent very recently that IL-6 is a major up-regulator of IGFBP-1 production by liver cells (26). Thus, not only is IL-6 itself subjected to regulation by chaperone proteins in the peripheral circulation, but also that IL-6 regulates the levels of proteins that chaperone other circulating growth factors. We anticipate that patients subjected to active anti-cancer immunotherapy regimens that result in elevations in IL-6 levels may also display altered IGFBP-1 regulation.
In the present study the detection in serum of high molecular mass
complexes of IL-6 by Sephadex gel filtration chromatography and the
variable reactivities of circulating IL-6 complexes in different ELISAs
led us to consider the roles of various IL-6 binding proteins in
regulating the bioavailability of this ligand in vivo. Potential IL-6
binding proteins include autoantibodies, sIL-6R (without or with the
additional association of sgp130), and other candidate proteins, such
as C-reactive protein and fragments of complement C3 and C4 (5, 6, 7, 8, 9, 10, 27, 28). That the human circulation consistently contains high levels of
IL-6 binding proteins in the form of sIL-6R (the soluble form of the
-chain of the cell surface receptor for IL-6; serum concentration,
10–100 ng/ml) (27, 28) and sgp130 (the soluble form of the ß-chain
of the cell surface receptor for IL-6; serum concentration,
300–400 ng/ml) (29) suggests that free IL-6 may at best have only a
transient existence in blood.
Anti-IL-6 IgG and anti-IL-6R IgG in the IL-6/sIL-6R-containing 200-kDa complexes from patients subjected to repeated polyclonal stimulation by BCG and KLH were experimentally demonstrated. These autoantibodies appear to generate reservoirs of IL-6 from which functional IL-6 may be released over a long period of time (6). Why AAAP vaccination leads to generation of the 30- and 450-kDa IL-6 complexes and not 200-kDa complexes is unclear and remains a subject for future investigation. However, it is now clear that different active immunization regimens lead to different abnormalities of IL-6 transport in blood.
Severe discrepancies in serum IL-6 data derived using different ELISAs for IL-6 were observed in the present study. None of the commercial ELISAs used in this study (R&D, Genzyme, and Endogen) had been previously evaluated for their ability to detect serum IL-6 in the face of IL-6 transport abnormalities in blood. Whether similar discrepancies occur in estimates of IL-6 levels in other body fluids, such as amniotic fluid, await evaluation.
Considerable effort is being expended by various investigators in modelling complexes of rIL-6, sIL-6R, and sgp130 as monovalent trimeric and divalent hexameric complexes (the latter of molecular mass 440 kDa) and an evaluation of the structural interactions and biologic properties of such complexes (reviewed in Refs. 20 and 30). These studies, conducted with the underlying premise that dimerization of gp130 leads to intracellular signaling, seek to identify structures that might have potent IL-6 antagonist or even superagonist activity using free monomeric IL-6 as the basis for comparison (30). The observation that IL-6 in blood exists in the form of differentially regulated high molecular mass complexes suggests that caution should be exercised in extrapolating the efficacy of a candidate IL-6 antagonist from cell culture to the in vivo situation.
Strategies to interfere with IL-6 function in man include administration of anti-IL-6 or anti-sIL-6R mAb (7, 31) or chimeric bipartite "ligand traps" consisting of an IgG Fc fragment dimer with one Fc portion covalently linked to the soluble portion of IL-6R and the other linked to the soluble portion of gp130 (32). The dichotomy between the properties of IL-6 antagonists in cell culture experiments and the properties observed in in vivo models point to a need to understand IL-6 transport in blood. As an example, in the case of anti-IL-6 mAb there is a paradoxical enhancement of IL-6 function in vivo despite the potent neutralizing properties of these mAb in cell culture experiments (6). It appears that administration of anti-IL-6 mAb leads to the generation in vivo of a long-lived intravascular pool of IL-6 of endogenously induced or exogenously administered origin that circulates as a 200-kDa complex bound to the IgG, from within which active cytokine can be released to target tissues, leading to an overall enhancement rather than an inhibition of cytokine function in vivo (6). As for the chimeric bipartite IL-6 ligand trap (32), its in vivo properties remain to be elucidated.
Active specific immunotherapy is increasingly in use as a treatment modality in human cancer. In many instances, the prerequisites for a therapeutic response remain unclear. In the case of immunotherapy of melanoma with the anti-Id mAb MK2–23 (such as the group II patients in this study), the appearance of an anti-anti-Id (Ab3) response and development of melanoma-specific cytotoxic T cells were positively correlated with an anti-tumor response (12, 13). The administration of BCG as adjuvant and the use of KLH as carrier together with the anti-Id mAb MK2–23 enhanced the Ab3 response (12, 13). In the case of AAAP immunotherapy, the predictors of therapeutic success are less clear. At a minimum, the present study shows that aggressive active immunization in humans leads to characteristic and dramatic alterations in blood IL-6 transport and bioavailability.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Dr. Pravin B. Sehgal, Department of Cell Biology and Anatomy, Basic Science Building, New York Medical College, Valhalla, NY 10595.
3 Abbreviations used in this paper: sIL-6R, soluble interleukin-6 receptor; KLH, keyhole limpet hemocyanin; BCG, Calmette-Guérin bacillus; AAAP, autologous anti-cancer antigen preparation; Vo, bed volume; Ve, excluded volume; Vi, included volume; sgp130, soluble gp130; IGFBP, insulin-like growth factor I binding protein; IGF-I, insulin-like growth factor I.
Received for publication June 16, 1997. Accepted for publication September 11, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  A. A. Beharka, M. Meydani, D. Wu, L. S. Leka, A. Meydani, and S. N. Meydani Interleukin-6 Production Does Not Increase With Age J. Gerontol. A Biol. Sci. Med. Sci., February 1, 2001; 56(2): 81B - 88. [Abstract] [Full Text]
|
|
|
|
 |
|
 R. S. Mummery and C. C. Rider Characterization of the Heparin-Binding Properties of IL-6 J. Immunol., November 15, 2000; 165(10): 5671 - 5679. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. M. Rollwagen, Y.-Y. Li, N. D. Pacheco, E. J. Dick, and Y.-H. Kang Microvascular Effects of Oral Interleukin-6 on Ischemia/Reperfusion in the Murine Small Intestine Am. J. Pathol., April 1, 2000; 156(4): 1177 - 1182. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. Dibbs, J. Thornby, B. G. White, and D. L. Mann Natural variability of circulating levels of cytokines and cytokine receptors in patients with heart failure: implications for clinical trials J. Am. Coll. Cardiol., June 1, 1999; 33(7): 1935 - 1942. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Muller-Newen, A. Kuster, U. Hemmann, R. Keul, U. Horsten, A. Martens, L. Graeve, J. Wijdenes, and P. C. Heinrich Soluble IL-6 Receptor Potentiates the Antagonistic Activity of Soluble gp130 on IL-6 Responses J. Immunol., December 1, 1998; 161(11): 6347 - 6355. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. J. Rayanade, M. I. Ndubuisi, J. D. Etlinger, and P. B. Sehgal Regulation of IL-6 Signaling by p53: STAT3- and STAT5-Masking in p53-Val135-Containing Human Hepatoma Hep3B Cell Lines J. Immunol., July 1, 1998; 161(1): 325 - 334. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and 4IL6/5IL6 (
)
ELISAs, calibrated using the World Health Organization 88/548
international standard for human IL-6. B, IL-6
concentrations assayed using three different commercial kits and
expressed in terms of the kit standards: R&D (
), Genzyme (
), and
Endogen (+). C, IL-6 concentrations assayed using the B9
hybridoma proliferation (•) or the hepatocyte Hep3B stimulation
bioassays (
), calibrated using the World Health Organization 89/548
standard for human IL-6.
