J Appl Physiol 98: 257–263, 2005.
First published July 30, 2004; doi:10.1152/japplphysiol.01090.2003.
Light chains are a ligand for megalin
R. Bryan S. Klassen,1,3 Patricia L. Allen,2,3 Vecihi Batuman,2,3
Kimberly Crenshaw,1 and Timothy G. Hammond2,3
1
Department of Chemistry, Xavier University, 2Tulane University School of Medicine, Tulane/Veterans Affairs
Environmental Astrobiology Center, and 3New Orleans Veterans Affairs Medical Center, New Orleans, Louisiana
Submitted 8 October 2003; accepted in final form 23 July 2004
in the urine of patients with multiple
myeloma may be nephrotoxic. They appear to exert their
effects by several independent mechanisms at different
nephron sites (7, 13, 14). Some nephrotoxic light chains can
produce glomerular damage (amyloidosis and light chain deposition disease) either in combination with tubular pathology or
alone (7, 13, 14). Most of the proximal tubular damage results
in an acute tubulopathy that histopathologically and clinically
mimics acute tubular necrosis (13, 14). A minority of patients
have Fanconi syndrome (i.e., glucosuria, aminoaciduria, and
phosphaturia) (13, 14). Those patients, although also showing
proximal tubular damage, also exhibit quite characteristic intracytoplasmic inclusions with a somewhat tubular/crystalline ultrastructural appearance (7, 13, 14). A specific abnormality, an unusual hydrophobic residue, in the pathologic
light chain in position 30 has been elucidated in patients
with Fanconi syndrome (14). It is important that acute
tubular necrosis and the changes associated with Fanconi
syndrome be properly differentiated for clinical and pathologic purposes (13).
The chronic tubulointerstitial nephropathy that commonly
accompanies multiple myeloma appears to be mediated by
cytokines released by proximal tubular cells as a result of
excessive light chain endocytosis (21). This cytokine response
appears to be mediated by MAP kinases (20).
Light chains of diverse origin bind to renal tubular proteins.
Mechanistic studies on renal tubular proteins demonstrate that
light chains, independent of their source, consistently bind to
cubilin in the proximal tubule (2) and to Tamm-Horsfall
protein in the distal tubule (11). Cubilin (gp280) is a giant
glycoprotein receptor and peripheral membrane protein, abundant in the clefts between apical brush-border membrane microvilli (2, 16, 18). Tamm-Horsfall protein is a glycoprotein
secretory product of the ascending limb of the loop of Henle (11).
Varied lines of evidence identify cubilin as an endocytic
receptor for some immunoglobulin light chains (2). First, light
chains coelute during immunoaffinity purification of cubilin.
Next, polyclonal antisera to cubilin, but not control sera,
displace human light chain binding from rat renal brush-border
membranes. Furthermore, anti-cubilin antiserum inhibit light
chain uptake into cubilin-expressing visceral yolk sac endothelial cells. Last, cubilin bind multiple species of light chains in
surface plasmon resonance (SPR) experiments. Binding shows
dose- and time-dependent saturability with low-affinity, highcapacity equilibrium binding parameters with dissociation constants in micromolar range (1, 2). These data demonstrate that
cubilin plays a role in the endocytosis and trafficking of light
chains in renal proximal tubule cells.
In addition to cubilin, however, several lines of evidence
converge to suggest that another abundant proximal tubular
glycoprotein receptor, megalin, also plays a role in light chain
endocytosis (23, 24). Classic binding kinetics studies using
Scatchard analysis demonstrate that a number of ligands compete with light chains for brush-border membrane binding.
These ligands include lysozyme, insulin, cytochrome c, myoglobin, and ␤2-microglobulin (2, 4, 25). Several of these
ligands have been shown elsewhere to be ligands for megalin
(4). Additionally, in conscious rats, lysozyme, cytochrome c,
and ␤2-microglobulin compete with each other for uptake in
the proximal tubules (3). Taken together, these observations
are consistent with the hypothesis that light chains bind to
megalin as well as cubilin as the entry step to proximal tubular
reabsorption. This study tests this hypothesis with both direct
molecular and membrane techniques.
The molecular structures of both cubilin and megalin are
consistent with multiple ligand binding domains of diverse
affinities. Megalin is a 600-kDa glycoprotein receptor with a
Address for reprint requests and other correspondence: R. B. Klassen, Dept.
of Chemistry, Box 114-B, Xavier Univ. of Louisiana, 1 Drexel Dr., New
Orleans, LA 70125-1098 (E-mail: [email protected]).
The costs of publication of this article were defrayed in part by the payment
of page charges. The article must therefore be hereby marked “advertisement”
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
cubilin; cancer; flow cytometry
LIGHT CHAINS PRESENT
http://www.jap.org
257
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
Klassen, R. Bryan S., Patricia L. Allen, Vecihi Batuman, Kimberly Crenshaw, and Timothy G. Hammond. Light chains are a
ligand for megalin. J Appl Physiol 98: 257–263, 2005. First published
July 30, 2004; doi:10.1152/japplphysiol.01090.2003.—Cubilin and
megalin are giant glycoprotein receptors abundant on the luminal
surface of proximal tubular cells of the kidney. We showed previously
that light chains are a ligand for cubilin. As cubilin and megalin share
a number of common ligands, we further investigated the ligand
specificity of these receptors. Three lines of evidence suggest that
light chains can also bind megalin: 1) anti-megalin antiserum largely
displaces brush-border light chain binding and megalin-expressing
BN-16 cell uptake more than anti-cubilin antiserum, 2) direct binding
studies on isolated proteins using surface plasmon resonance techniques confirm that megalin binds light chains, and 3) light chains
compete with known megalin ligands for brush-border membrane
binding and BN-16 cell uptake. The megalin-light chain interaction is
divalent ion dependent and similar for both ␬- and ␭-light chains. A
fit of the data on light chain binding to megalin over a concentration
range 0.078 –2.5 mg/ml leads to an estimated dissociation constant of
6 ⫻ 10⫺5 M, corresponding approximately to one light chain-binding
site per megalin and in the same range for dissociation constants for
cubilin binding. These data suggest that light chains bind the tandem
megalin-cubilin complex. Megalin is the major mediator of light chain
entry into megalin-expressing membrane such as the apical surface of
proximal tubular epithelial cells.
258
MYELOMA LIGHT CHAINS BIND MEGALIN
single transmembrane domain (15, 19). It attracted intense
interest after it was identified as the target of glomerulonephritic antisera, becoming known as “the Heymann antigen”
(6, 15, 19). Saito et al. (19) cloned the protein, finding a
5,044-residue protein with high homology to the low-density
lipoprotein receptor family. The structure of cubilin, a 460-kDa
protein, is characterized by 27 consecutive low-affinity ligand
binding domains (2, 4), followed by 8 EGF repeats, but no
transmembrane span (16). Cubilin is a peripheral membrane
protein anchored to the apical membrane of proximal tubule
cells via attachment to megalin (16).
METHODS AND MATERIALS
J Appl Physiol • VOL
RESULTS
We compared displacement of fluorescent light chain binding to renal brush-border membrane vesicles by various anti-
98 • JANUARY 2005 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
Animals, reagents, and antibodies. Male Sprague-Dawley rats
(200 –250 g) were obtained from Sasco (Omaha, NE). All other
reagents were from Sigma Chemicals (St. Louis, MO) unless otherwise stated.
Polyclonal antibodies against cubilin and megalin were raised in
rabbits against proteins purified by immunoaffinity chromatography
using previously reported monoclonal antibodies coupled to Sepharose 4B (2, 10, 17, 18). These antibodies, used as whole serum
preserved with azide, were monospecific by immunoblotting on whole
brush border preparations and by immunoprecipitation of biosynthetically labeled yolk sac epithelial cells in culture (17, 18). For a control
antiserum, available antisera unrelated to cubilin or megalin but of the
same species and isotype were surveyed by flow cytometry to find a
serum with the same peak binding as the cubilin and megalin antisera.
An anti-neurokinin (NK)-1/substance P receptor antiserum (a kind gift
of Dr. Jacques Couraud, Gif-sur-Yvette, France) proved suitable (5).
Light chains. To study a defined group of light chains, we selected
light chains from patients with pure light chain proteinuria without
albuminuria, consistent with a lack of glomerular disease. The patients
did not have Fanconi syndrome clinically, but they did have tubulointerstitial nephropathy, consistent with distal cast formation. The
light chains, purified as previously discussed (2, 3, 9), were monomeric on Western blot and showed a single band of appropriate size
on an SDS-PAGE gel. Two ␬-light chains and two ␭-light chains were
studied.
Preparation of rat brush-border membrane vesicles, antibody binding studies, and ligand interference studies. Rat renal cortical brushborder membrane vesicles were isolated by magnesium precipitation
techniques as described previously (1, 2, 10). Human ␭-light chains
were labeled with FluorX, Cyanine-3 (Amersham, Boston, MA), or
FITC (Sigma Chemical). Residual free dye was separated from light
chain conjugated dye by dialysis into a suitable binding buffer using
a Slidalyzer cassette (Pierre, Rockford, IL). Binding of light chains
was investigated in presence of 1,000- to 2,000-fold dilutions of
anti-cubilin or anti-megalin polyclonal antibodies recognizing the
holoprotein (2, 10, 17, 18). Binding of FluorX conjugated light chains
to brush-border membrane vesicles, with and without antibody or
ligand interference, was analyzed by flow cytometry using a FACStar
Plus flow cytometer to collect data files of 2,000 observations per
sample. FluorX was excited with the 488 nm line of an argon-ion
laser, and fluorescence units were measured at 530 ⫾ 30 nm using a
photo multiplier and log amplification of the arbitrarily zeroed signals
(2, 10). Data files were collected and analyzed using CellQuest
software. For ligand interference studies, one light chain was fluorescent labeled and competition with other light chains assayed (2).
Cyanine-3 conjugated metallothionein (MT) was prepared using the
same method as for light chains.
Direct Biacore analysis of light chain/cubilin interaction using
surface plasmon resonance technology. Surface plasmon resonance
(SPR) experiments used a BIACORE 3000 instrument. Immobilizations were conducted at 25°C with a flow buffer at pH 7.4 containing
10 mM HEPES, 2 mM CaCl2, 150 mM NaCl, and 0.005% NP-40. A
Biacore CM5 dextran sensor chip was activated with N-hydroxysuccinimide and N-ethyl carbodiimide according to manufacturer protocols and cubilin (in 10 mM acetate, pH 4.8) was injected immediately
afterward. Unreacted sites were blocked with 1 M ethanolamine (pH
8.5) (9, 16, 20, 22). When a control (reference) surface was needed for
quantitative measurements, an equal amount of casein (0.20 mg/ml in
10 mM acetate, pH 4.78) was immobilized in a parallel flow cell.
Surface plasmon resonance study of the dose-dependent binding of
myeloma light chains to megalin. A solution of myeloma light chains
(3.4 mg/ml in HBS) was serially diluted to prepare solutions ranging
from 0.078 to 2.5 mg/ml. The blank contained only buffer. These
solutions were repeatedly injected over the surface for 2.2 min (110 ␮l
at 50 ␮l/min, 25°C). The association (injection) phase and a subsequent 5-min dissociation phase were monitored and corrected in
real-time by subtracting the reference signal. After several minutes,
the light chains had not entirely dissociated from the megalin, so a
mildly acidic cocktail (pH 5.0) was injected to regenerate the surface.
The raw binding data for the megalin cell were corrected by subtracting the casein reference signal to give the net response. Binding
kinetics including association and dissociation constants were analyzed using Biacore analysis software.
Cell culture studies. Except as noted, experiments were conducted
using immortalized yolk sac cells from the Brown Norway rat (BN16) (17). An apical brush border and a specialized endosomal pathway
similar to the renal proximal tubule, including abundant expression of
megalin and cubilin, characterize these cells. The cells were grown in
DMEM (Gibco/Invitrogen, Carlsbad, CA) supplemented with 10%
fetal calf serum and 50 ␮g/ml streptomycin or ciprofloxacin. Cells
were passaged every 4 days with a split ratio of 10:1. Madin-Darby
canine kidney cells were grown in a modified minimal essential
medium as described in ATCC (Manassas, VA) protocols.
Light chain uptake by BN-16 cells analyzed by flow cytometry.
Uptake experiments used FluorX- or Cy3-labeled light chains and
were performed with confluent monolayers cultured in 96-well plates.
In preliminary experiments, we determined that light chain uptake was
linear for at least 3 h and exhibited dose-dependent saturation. The
concentration producing half-maximal uptake was ⬃5 ␮M. Inhibition
experiments were performed as follows. The confluent monolayers
were washed with serum-free DMEM and allowed to equilibrate for
2 h at 37°C. The cells were then incubated with 5 ␮M labeled light
chains and any inhibitor for 1–2.5 h at 37°C. Incubations were
performed in DMEM containing 0.1% ovalbumin to reduce nonspecific binding. The cells were washed several times with PBS, acidwashed to release membrane-bound proteins, released with trypsin,
and washed several more times with PBS. In this state they could be
analyzed immediately, without fixing, by flow cytometry analysis.
The positive control was labeled light chains without added inhibitor;
the negative control was unlabeled light chains. Inhibitor concentrations were generally 10 –100 times greater than the concentration of
labeled light chains.
Statistics. Data are expressed as means ⫾ SE throughout the
manuscript. Statistical analysis was performed by analysis of variance
and Scheffé’s post hoc comparison. The large data sets of 2,000
individual observations generated by flow cytometry analysis were
analyzed for statistical significance by Kolgomorov-Smirnov summation statistics (27). Flow cytometry measurements are log amplified to
allow collection of the typically broad range of fluorescent values on
a defined scale. Both the arithmetic and geometric means of the 2,000
individual cell measurements collected are calculated, along with the
coefficient of variation (CV). If the distribution is tightly Gaussian,
such that the CV is ⬍0.5 the mean, then the arithmetic mean is shown,
but if CV ⬎ 0.5 is the mean then the geometric correction is used.
MYELOMA LIGHT CHAINS BIND MEGALIN
Fig. 1. Displacement of light chain binding to rat renal cortical brush-border
membranes by anti-cubilin and megalin antisera using flow cytometry analysis.
Anti-cubilin polyclonal antiserum inhibits FITC-light chain binding to rat renal
brush-border membrane vesicles, but not nearly as much as megalin antiserum.
In contrast, both normal rabbit serum and antisera to the neurokinin-1/
substance P receptor, which binds these membranes, has no effect on light
chain binding. This suggests the cubilin and megalin antiserum effects are
highly specific. Data are expressed as means ⫾ SE with statistical analysis by
analysis of variance and Scheffé’s post hoc comparison.
J Appl Physiol • VOL
Fig. 2. Flow cytometry analysis of the displacement of fluorescently labeled
light chains (LC) from brush-border membranes by known megalin ligands.
Brush-border membrane vesicles isolated from rat renal cortex were incubated
with fluorescently conjugated LC and various known megalin ligands. A:
observed fluorescence is shown for the control (no LC) and fluorescent LC
alone (FITC-LC), as well as the effect of 100⫻ competing unlabeled LC
(FITC-LC ⫹ LC) or FITC-LC and 100⫻ myoglobin (FITC-LC⫹myoglobin).
B: Cy3 conjugated metallothionein (Cy3-MT) was displaced by both unlabeled
LC (Cy3-MT ⫹ LC) and myoglobin (Cy3-MT ⫹ myoglobin). Data files of
2,000 observations per sample were collected, n ⫽ 3, mean ⫾ SE shown, with
statistical analysis by Kolgomorov-Smirnov summation statistics.
rescent units, n ⫽ 3, P ⬍ 0.05 on each run by KolgomorovSmirnov summation statistics). The known megalin ligand,
MT, fluorescently labeled with cyanine-3 (Cy3) bound significantly to brush-border membrane vesicles at 3 ␮M concentration (control unlabeled 4 ⫾ 1 fluorescent units, mean ⫾ SE,
labeled 41 ⫾ 5, n ⫽ 4, P ⬍ 0.05; Fig. 2B). Unlabeled ␬-1 light
chains at 30 ␮M inhibited Cy3-MT binding (17 ⫾ 2 fluorescent
units, n ⫽ 3, P ⬍ 0.05 on each run by Kolgomorov-Smirnov
summation statistics), as did the known megalin ligand myoglobin at 30 ␮M (30 ⫾ 5 fluorescent units, n ⫽ 3, P ⬍ 0.05 on
each run by Kolgomorov-Smirnov summation statistics).
Independent confirmation of the flow cytometry results was
obtained through Biacore SPR measurements, which also
showed that light chains bind to megalin (Fig. 3). SPR analyses
allow for direct real-time binding assays between an immobilized molecule and its soluble ligand (or its receptor, depending
on the experimental design) (12). In the case of light chains
exposed to immobilized megalin, the responses for multiple
injections of each light chain sample [light chains ␬-1, ␬-2, and
␭-1 were highly reproducible over the concentration range
0.078 –2.5 mg/ml (Fig. 3A)]. At all concentrations of light
chain, the signal reached near equilibrium within 2 min. A fit
of these data leads to an estimated dissociation constant of 6 ⫻
10⫺5 M, a value similar to previously observed constants from
our laboratory (1) (Fig. 3B). Additionally, these data indicate a
maximum binding of ⬃75 response units (RU), corresponding
approximately to one light chain binding site per megalin. With
98 • JANUARY 2005 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
sera using flow cytometry techniques. In preliminary experiments we observed that anti-megalin and anti-cubilin antisera
had classic binding curves to brush-border membrane vesicles
and had the same peak titer. We surveyed a number of
irrelevant sera of matching animal host and determined that an
antibody to the NK-1 receptor had the same peak binding
dilution as the anti-megalin and anti-cubilin sera, making it a
suitable control. Human FITC-conjugated ␬-1 light chains
(FITC-LC) were used in this experiment. FITC-LC bound to
rat renal brush-border membrane vesicles were displaced by
polyclonal antibodies to either cubilin or megalin (Fig. 1).
Initial light chain binding (45.5 ⫾ 4.3 fluorescent units, n ⫽ 8)
increased compared with unstained membranes (5.1 ⫾ 1.2
units, n ⫽ 8; P ⬍ 0.05). Subsequently, bound light chains were
displaced by anti-cubilin (30.2 ⫾ 1.0 units, n ⫽ 8; P ⬍ 0.05),
but far more potent displacement was observed with antimegalin antisera (12.9 ⫾ 2.8 units, n ⫽ 8; P ⬍ 0.05). In
contrast, normal rabbit serum (NRS; 42.9 ⫾ 1.7 units, n ⫽ 8)
and antisera to the NK-1/substance P receptor (40.0 ⫾ 1.2
units, n ⫽ 4) produced no effect. This suggests that the
anti-cubilin and anti-megalin antiserum effects are highly specific. This flow cytometry data provide the first line of evidence
that megalin, like cubilin, is a receptor for light chains, and, in
fact, megalin appears to bind more light chains than cubilin.
Next we tested for competition between light chains and
known megalin ligands for brush-border membrane vesicle
binding (Fig. 2). Light chain ␬-1 was used for these experiments. Light chains fluorescently labeled with FITC bound
significantly to brush-border membrane vesicles at 3 ␮M
concentration (control unlabeled 5 ⫾ 1 fluorescent units,
mean ⫾ SE, labeled 374 ⫾ 42, n ⫽ 4, P ⬍ 0.05; Fig. 2A).
Unlabeled ␬-1 light chains at 30 ␮M inhibited FITC-LC
binding (83 ⫾ 3 fluorescent units, n ⫽ 3, P ⬍ 0.05 on each run
by Kolgomorov-Smirnov summation statistics), as did the
known megalin ligand myoglobin at 30 ␮M (119 ⫾ 14 fluo-
259
260
MYELOMA LIGHT CHAINS BIND MEGALIN
a minor sigmoidal element in the early part of the binding
curves we cannot exclude a more complex cooperativity
model, but this would not change the conclusions regarding
megalin/light chain interactions proposed from this analysis.
As previously reported (2), the affinity of light chains for
cubilin is in micromolar range and similar to the megalin/light
chain interaction observed here.
Light chain uptake in cultured cells and inhibition by antibodies and ligands studied by flow cytometry. To determine the
cellular uptake of fluorescently labeled light chains and inhibition by known megalin ligands we began with dose- and
time-dependent uptake studies. Light chain ␬-1 was used in
these experiments. Incubation of BN-16 cells with 0 – 80 ␮M
fluorescently labeled light chains for 3 h followed by flow
cytometry analysis demonstrates that light chain uptake is
saturable and that light chain concentrations of 4 –5 ␮M produce half-maximal uptake (data not shown). Uptake of light
chains by BN-16 cells was then demonstrated to be linearly
time dependent at doses above and below the half-maximal
binding concentration (data not shown).
Incubating BN-16 cells with anti-megalin antibodies before
adding 10 ␮M fluorescent light chain greatly reduced the
uptake of light chain in a dose-dependent manner (Fig. 4)
(unstained cells 8 ⫾ 3 arithmetic mean ⫾ SE, increases to
1,374 ⫾ 141; n ⫽ 3, P ⬍ 0.05; with anti-megalin antisera at
1:100 dilution, 202 ⫾ 13; at 1:330 dilution, 515 ⫾ 12; at
1:1,000 dilution, 823 ⫾ 168; at 1:3,300 dilution, 1,106 ⫾ 36;
n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov summation statistics
on each run). At the same titer, anti-cubilin antiserum had a
smaller, although significant, effect (at 1:100 dilution, 889 ⫾
J Appl Physiol • VOL
65; at 1:330 dilution, 1,312 ⫾ 69; at 1:1,000 dilution, 1,195 ⫾
138; at 1:3,300 dilution, 1,366 ⫾ 74; n ⫽ 3, P ⬍ 0.05 by
Kolgomorov-Smirnov summation statistics on each run for
dilutions 1:100, 1:330, and 1:1,000). Angiotensin II type 1
(AT1) receptor antiserum, in contrast, had no effect (at 1:100
dilution, 1,255 ⫾ 46; at 1:330 dilution, 1,245 ⫾ 35; at 1:1,000
dilution, 1,245 ⫾ 101; at 1:3,300 dilution, 1,182 ⫾ 124; n ⫽
3). Antibodies against megalin and cubilin had an additive
Fig. 4. Flow cytometry analysis of antibody inhibition of uptake of fluorescently labeled MT into BN-16 cells. Anti-cubilin antisera inhibited MT uptake
into BN-16 cells in a concentration-dependent manner. Effect of anti-megalin
antisera was far greater than anti-cubilin antisera; the 2 sera produced an
additive effect. Anti-AT1-receptor antiserum, which also binds BN-16 cells,
was used as a nonspecific binding control but had no effect on MT uptake. Data
files of 2,000 observations per sample were collected, n ⫽ 3, mean ⫾ SE
shown, with statistical analysis by Kolgomorov-Smirnov summation statistics.
98 • JANUARY 2005 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
Fig. 3. Dose-dependent binding of myeloma
LC to megalin using surface plasmon resonance (SPR) analysis. Myeloma LC (3.4 mg/
ml) was serially diluted in HBS. These solutions were injected 3 times over the surface
for 2.2 min. A: from bottom to top, the traces
represent the corrected real-time responses
obtained for concentrations of 0, 0.078,
0.156, 0.312, 0.625, 1.25, and 2.5 mg/ml. B:
fit of the maximum (equilibrium) responses
obtained within 2 min.
MYELOMA LIGHT CHAINS BIND MEGALIN
261
Fig. 5. Flow cytometry analysis of competition between
LC for BN-16 cell uptake. Concentration-dependent
competition between LC for BN-16 cell uptake. Data
files of 2,000 observations per sample were collected,
n ⫽ 3, mean ⫾ SE, with statistical analysis by Kolgomorov-Smirnov summation statistics.
J Appl Physiol • VOL
competed with fluorescent light chain (341 ⫾ 26 at 10 ␮M,
260 ⫾ 51 at 25 ␮M and 175 ⫾ 9 at 60 ␮M, n ⫽ 3, P ⬍ 0.05
by Kolgomorov-Smirnov summation statistics on each run), as
did retinol binding protein (330 ⫾ 86 at 10 ␮M, and 96 ⫾ 9 at
35 ␮M, n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov summation
statistics on each run). MT also competed for light chain
binding (181 ⫾ 34 at 10 ␮M, 217 ⫾ 90 at 35 ␮M and 134 ⫾
6 at 100 ␮M, n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov
summation statistics on each run).
The pattern was quite similar when fluorescent light chain
was used at 9 ␮M rather than 4 ␮M (Fig. 6B). There was
significant uptake of 4 ␮M light chain at 2 h (control 12 ⫾ 2
fluorescent units, mean ⫾ SE compared with light chain 598 ⫾
Fig. 6. Flow cytometry analysis of competitive displacement of uptake of
fluorescently labeled LC into BN-16 cells by known megalin ligands. A:
competition of the known megalin ligands ␤2-microglobulin, retinol binding
protein, and apolipoprotein A-1 with MT for BN cell uptake using 4 ␮M LC.
B: same as A but with 9 ␮M LC. Data files of 2,000 observations per sample
were collected, n ⫽ 3, mean ⫾ SE, with statistical analysis by KolgomorovSmirnov summation statistics.
98 • JANUARY 2005 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
effect (at 1:100 dilution, 68 ⫾ 1; at 1:330 dilution, 229 ⫾ 25;
at 1:1,000 dilution, 594 ⫾ 132; at 1:3,300 dilution, 1,088 ⫾ 26;
n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov summation statistics
on each run against both anti-megalin and anti-cubilin alone).
The effect of antibodies on light chain uptake was not observable unless ovalbumin was used to reduce nonspecific effects.
To determine if there is competition between light chains for
BN-16 cell uptake, 3 ␮M fluorescently labeled ␬-1 light chain
was added with 3, 30, or 300 ␮M of the four light chains, ␬-1,
␬-2, ␭-1, and ␭-2, and fluorescent light chain uptake assayed by
flow cytometry (Fig. 5). Compared with unlabeled cells (11 ⫾
1 fluorescent units, mean ⫾ SE, n ⫽ 3), fluorescent light chains
were taken up into the BN-16 cells after 1 h of incubation
(107 ⫾ 11 fluorescent units, n ⫽ 3, P ⬍ 0.05 for each run by
Kolgomorov-Smirnov summation statistics). Each of the four
light chains competed for uptake: ␬-1 reduced fluorescent light
chain uptake as follows: 155 ⫾ 34 fluorescent units at 3 ␮M,
99 ⫾ 4 at 30 ␮M, and 75 ⫾ 6 at 300 ␮M (n ⫽ 3, P ⬍ 0.05 for
each run by Kolgomorov-Smirnov summation statistics). At
the lowest 3 ␮M concentration of this light chain, aggregates
could be seen in the size parameters on the flow cytometer,
which likely accounts for the apparent increase in uptake over
baseline. Similarly, ␬-2 permitted 94 ⫾ 13 fluorescent units at
3 ␮M, 68 ⫾ 51 at 30 ␮M, and 34 ⫾ 25 at 300 ␮M (n ⫽ 3, P ⬍
0.05 for each run by Kolgomorov-Smirnov summation statistics). ␭-1 permitted 78 ⫾ 8 fluorescent units at 3 ␮M, 55 ⫾ 4
at 30 ␮M, and 36 ⫾ 2 at 300 ␮M (n ⫽ 3, P ⬍ 0.05 for each
run by Kolgomorov-Smirnov summation statistics). ␭-2 permitted 94 ⫾ 10 fluorescent units at 3 ␮M, 83 ⫾ 23 at 30 ␮M,
and 70 ⫾ 9 at 300 ␮M (n ⫽ 3, P ⬍ 0.05 for each run by
Kolgomorov-Smirnov summation statistics).
The addition of known megalin ligands ␤2-microglobulin,
retinol binding protein, and apolipoprotein A-1, reduced ␬-1
light chain uptake in a dose-dependent manner across a broad
range of concentrations of both light chain and the other
ligands (Fig. 6). Light chains showed competition for their own
binding sites when Cy3 fluorescently conjugated MT was
incubated with nonfluorescent light chain, cell washed in acid
to remove surface binding, trypsinized off a culture plate, and
analyzed by flow cytometry. There was significant uptake of 4
␮M light chain at 2 h (control 12 ⫾ 2 fluorescent units,
mean ⫾ SE compared with light chain 409 ⫾ 60, n ⫽ 3, P ⬍
0.05 by Kolgomorov-Smirnov summation statistics on each
run; Fig. 6A). Unlabeled light chain competed with fluorescent
apolipoprotein A-1 (80 ⫾ 23 at 10 ␮M, and 27 ⫾ 9 at 35 ␮M
n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov summation statistics
on each run). The known megalin ligand ␤2-microglobulin
262
MYELOMA LIGHT CHAINS BIND MEGALIN
85, n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov summation
statistics on each run). Unlabeled light chain competed with
fluorescent apolipoprotein A-1 (94 ⫾ 14 at 10 ␮M, and 22 ⫾
4 at 35 ␮M n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov
summation statistics on each run). The known megalin ligand
␤2-microglobulin competed with fluorescent light chain
(583 ⫾ 162 at 10 ␮M, n ⫽ 3 not significant, but 372 ⫾ 35 at
35 ␮M, n ⫽ 3, P ⬍ 0.05 by Kolgomorov-Smirnov summation
statistics on each run), as did retinol binding protein (283 ⫾ 33
at 10 ␮M and 222 ⫾ 74 at 35 ␮M, n ⫽ 3, P ⬍ 0.05 by
Kolgomorov-Smirnov summation statistics on each run). MT
also competed for LC binding (314 ⫾ 85 at 10 ␮M, 248 ⫾ 49
at 35 ␮M and 217 ⫾ 22 at 100 ␮M, n ⫽ 3, P ⬍ 0.05 by
Kolgomorov-Smirnov summation statistics on each run).
DISCUSSION
J Appl Physiol • VOL
GRANTS
This work was supported by National Institutes of Health (NIH) First
Award DK-46117 (T. G. Hammond), NIEHS ARCH ES-09996 (R. B. S.
Klassen/T. G. Hammond), National Aeronautics and Space Administration
Grant 9 – 811 Basic (T. G. Hammond), NASA Grant NAG8 –1362, a Veterans
Administration Research Associate Career Development Award (T. G. Hammond), Veterans Affairs Minority Research Program (T. G. Hammond/R. B.
Klassen), and Veterans Affairs Merit Review Funds (V. Batuman). Flow
cytometry equipment was purchased under the auspices of a Louisiana Education Quality Support Fund of the Board of Regents support fund grant. Work
was also supported in part by NIH Grant Number P20-RR-017659 from the
Institutional Development Award Program of the National Center for Research
Resources and by Louisiana Board of Regents through the Millennium Trust
Health Excellence Fund contract number HEF(2001– 06)-07.
REFERENCES
1. Batuman V, Driesbach AW, and Cyran J. Light-chain binding sites on
renal brush border membranes. Am J Physiol Renal Fluid Electrolyte
Physiol 258: F1259 –F1265, 1990.
2. Batuman V, Verroust PJ, Simon E, Pontillon F, Lyles M, Bruno J, and
Hammond TG. Myeloma light chains bind gp280 in the kidney. Am J
Physiol Renal Physiol 275: F246 –F254, 1998.
3. Bernard A, Amor AO, Viau C, and Lauwerys R. The renal uptake of
proteins: a nonselective process in conscious rats. Kidney Int 34: 175–185,
1988.
4. Bork B and Beckman G. The Domain: A CUB widespread module in
developmentally regulated proteins. J Mol Biol 231: 539 –545, 1993.
5. Bret-Dibat JL, Zouaoui D, Dery O, Zerari F, Grassi J, Maillet S,
Conrath M, and Couraud JY. Antipeptide polyclonal antibodies that
recognize a substance P-binding site in mammalian tissues: a biochemical
and immunocytochemical study. J Neurochem 63: 333–343, 1994.
6. Christensen EI, Birn H, Verroust P, and Moestrup SK. Membrane
receptors for endocytosis in the renal proximal tubule. Int Rev Cytol 180:
237–284, 1998.
7. Clark AD, Shetty A, and Soutar R. Renal failure and multiple myeloma:
pathogenesis and treatment of renal failure and management of underlying
myeloma. Blood Rev 13: 79 –90, 1999.
8. Deret S, Denoroy L, Lamarine M, Vidal R, Mougenot B, Frangione B,
Stevens FJ, Ronco PM, and Aucouturier P. Kappa light chain-associated Fanconi’s syndrome: molecular analysis of monoclonal immunoglobulin light chains from patients with, and without intracellular crystals.
Protein Eng 12: 363–369, 1999.
9. Driesbach AW and Batuman V. Low-molecular weight protein competition for binding sites in renal brush border membranes. Renal Physiol 17:
87–293, 1994.
10. Hammond TG, Majewski RR, Kaysen JH, Goda FO, Navar G,
Pontillon F, and Verroust PJ. Gentamicin inhibits rat renal cortical
98 • JANUARY 2005 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
Previous studies from our laboratories provided evidence
that nephrotoxic myeloma light chains bind cubilin (2). Two
␬-light chains and two ␭-light chains bound cubilin in a
dose-dependent, temperature-dependent, competitive manner
during SPR analysis of isolated protein binding. Antiserum to
cubilin abolished substantial amounts of the brush-border
membrane binding of a nephrotoxic light chain. Cubilin and
light chains coeluted from an anti-cubilin monoclonal affinity
column. Hence we concluded that myeloma light chains bind
cubilin (2). In the kidney, cubilin is restricted in distribution to
the proximal tubule, and, clinically, the proximal tubular damage of myeloma light chains results in either an acute tubulopathy that mimics acute tubular necrosis in the majority of
cases, whereas a minority of patients have Fanconi syndrome
(7, 13, 14). It is unclear to what extent the interaction of cubilin
with light chains mediates the clinical pathology of myeloma.
As light chains are the variable regions of an antibody they
are structurally diverse, and the critical determinants of proximal and distal toxicity have yet to be definitively identified (2,
11, 26). Excessive endocytosis and protein overloading in
proximal tubular cells appear to result in release of inflammatory cytokines that may be responsible for chronic tubulointerstitial nephritis, which is common in multiple myeloma (11,
20). These data suggest that light chains bind the tandem
megalin-cubilin complex, and some light chains bind preferentially to megalin rather than cubilin. Classic binding kinetics
studies using Scatchard analysis demonstrate that a number of
the ligands, which compete with light chains for brush-border
membrane binding, are known ligands for megalin (25). This
study provides direct surface plasmon resonance, membrane
vesicles antibody displacement data, and BN-16 cell uptake
data consistent with the finding that light chains are indeed a
ligand for megalin. The affinity of the light chains studied for
megalin is quite similar to the affinity of light chains for cubilin
and to affinity constants derived from radiolabeled light chain
binding to rat renal brush-border membranes previously observed in our laboratory (1).
The new insight to pathogenesis deduced from the current
studies is that even in patients without Fanconi syndrome, or
glomerular injury, but with clinical tubulointerstitial nephritis,
light chains bind megalin and are taken up by megalin-expressing cell epithelia. Hence it appears that megalin binding and
uptake are not sufficient for Fanconi syndrome formation or
glomerulopathy. Resistance to cathepsin B proteolysis and
subsequent accumulation in proximal tubular endosomes has
been associated with Fanconi Syndrome.
Messiaen et al. (14) reappraised the unifying concept that
patients with Fanconi syndrome and Bence Jones proteinuria
have a special form of plasma cell dyscrasia characterized by
slow progression of the tumor and by prominent crystal formation in proximal tubule cells, in the absence of myeloma
casts in the distal tubule (8). Resistance of the variable portion
of ␬-light chains to proteolysis in Fanconi syndrome patients
can explain the accumulation of the light chain in the endocytotic compartment of the proximal tubule cells, leading to impairment of proximal tubule functions.
This new understanding of the interactions of light chains
with the renal proximal tubular scavenger pathway receptors,
cubilin and megalin, demonstrates that megalin is the major
mediator of proximal tubular light chain reabsorption: the renal
cellular entry mechanism for light chains has now been defined. This new information may well prove important for the
eventual development of clinical protective agents for the
proximal tubular elements of the nephrotoxic damage caused
by light chains in patients with myeloma.
MYELOMA LIGHT CHAINS BIND MEGALIN
11.
12.
13.
14.
15.
16.
18.
J Appl Physiol • VOL
19.
20.
21.
22.
23.
24.
25.
26.
27.
chemistry and ontogeny of two closely related coated pit proteins—the
280-kd target of teratogenic antibodies and the 330-kd target of nephritogenic antibodies. Am J Pathol 142: 1654 –1667, 1993.
Saito A, Pietromonaco S, Loo AKC, and Farquhar MG. Complete
cloning and sequencing of rat gp330/“megalin,” a distinctive member of
the low density lipoprotein receptor gene family. Proc Natl Acad Sci USA
91: 9725–9729, 1994.
Sengul S, Zwizinski C, and Batuman V. Role of MAPK pathways in
light chain-induced cytokine production in human proximal tubule cells.
Am J Physiol Renal Physiol 284: F1245–F1254, 2003.
Sengul S, Zwizinski C, Simon EE, Kapasi A, and Singhal PC. Endocytosis of light chains induces cytokines through activation of NF-␬B in
human proximal tubule cells. Kidney Int 62: 1977–1988, 2002.
Van der Merwe PA and Barclay AN. Analysis of cell-adhesion molecule
interactions using surface plasmon resonance. Curr Opin Immunol 8:
257–261, 1996.
Verroust PJ, Birn H, Nielsen R, Kozyraki R, and Christensen EI. The
tandem endocytic receptors megalin and cubilin are important proteins in
renal pathology. Kidney Int 62: 745–756, 2002.
Verroust PJ and Kozyraki R. The roles of cubilin and megalin, two
multiligand receptors, in proximal tubule function: possible implication in
the progression of renal disease. Curr Opin Nephrol Hypertens 10: 33–38,
2001.
Verroust PJ, Kozyraki R, Hammond TG, Moestrup SK, and Christensen EI. Physiopathologic role of cubilin, and megalin. Adv Nephrol
Necker Hosp 30: 127–145, 2000.
Winearls CG. Nephrology forum: acute myeloma kidney. Kidney Int 48:
1347–1360, 1995.
Young IT. Proof without prejudice: use of the Kolgomorov-Smirnov test
for the analysis of histograms from flow systems and other sources.
J Histochem Cytochem 25: 935–941, 1977.
98 • JANUARY 2005 •
www.jap.org
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on June 18, 2017
17.
homotypic endosomal fusion: role of megalin. Am J Physiol Renal Physiol
272: F117–F123, 1997.
Huang ZQ and Sanders PW. Localization of a single binding site for
immunoglobulin light chains on human Tamm-Horsfall glycoprotein.
J Clin Invest 99: 732–736, 1997.
Jonsson U, Fagerstam L, Ivarsson B, Johnsson B, Karlsson R, Lundh
K, Lofas S, Persson B, Roos H, and Ronnberg I. Real-time biospecific
interaction analysis using surface plasmon resonance, and a sensor chip
technology. Biotechniques 11: 620 – 627, 1991.
Markowitz GS. Dysproteinemia and the kidney. Adv Anat Pathol 11:
49 – 63, 2004.
Messiaen T, Deret S, Mougenot B, Bridoux F, Dequiedt P, Dion JJ,
Makdassi R, Meeus F, Pourrat J, Touchard G, Vanhille P, Zaoui P,
Aucouturier P, and Ronco PM. Adult Fanconi syndrome secondary to
light chain gammopathy. Clinicopathologic heterogeneity and unusual
features in 11 patients. Medicine (Baltimore) 79: 135–154, 2000.
Moestrup SK, Cui S, Vorum H, Bregengaard C, Bjørn SE, Norris K,
Gliemann J, and Christensen EI. Evidence that epithelial glycoprotein
330/megalin mediates uptake of polybasic drugs. J Clin Invest 96: 1404 –
1413, 1995.
Moestrup SK, Kozyraki R, Christiansen M, Kaysen JH, Rasmussen
HH, Brault D, Pontillon F, Goda FO, Christensen EI, Hammond TG,
and Verroust PJ. The intrinsic factor-vitamin B12 receptor, and target of
teratogenic antibodies is a megalin-binding peripheral membrane protein
with homology to developmental proteins. J Biol Chem 273: 5235–5242,
1998.
Sahali D, Mulliez N, Chatelet F, Dupius R, Ronco P, and Verroust P.
Characterization of a 280-kD protein restricted to the coated pits of the
renal brush border, and the epithelial cells of the yolk sac. J Exp Med 167:
213–218, 1988.
Sahali D, Mulliez N, Chatelet F, Laurentwinter C, Citadelle D, Sabourin JC, Roux C, Ronco P, and Verroust PJ. Comparative immuno-
263