Membranous Glomerulonephritis Associated with Industrial
Mercury Exposure
Study of Pathogenetic Mechanisms
RAYMOND R. TUBBS, D.O., GORDON N. GEPHARDT, M.D., JAMES T. McMAHON, PH.D., MARC C. POHL, M.D.,
DONALD G. VIDT, M.D., SUMNER A. BARENBERG, PH.D., AND RAFAEL VALENZUELA, M.D.
Tubbs, Raymond R., Gephardt, Gordon N., McMahon, James
T., Pohl, Marc C, Vidt, Donald G., Barenberg, Sumner A.,
and Valenzuela, Rafael: Membranous glomerulonephritis associated with industrial mercury exposure. Study of pathogenetic mechanisms. Am J Clin Pathol 77: 409-413, 1982. The
nephrotoxicity associated with mercury may be manifested as
either acute tubular necrosis or an immune complex glomerulonephritis, depending upon the conditions under which the
patient is exposed to the metal. Two patients with industrial
exposure to mercury developed the nephrotic syndrome due to
membranous glomerulonephritis. A multidisciplinary approach
was used to define more precisely the pathogenetic mechanisms
involved in the production of the glomerular lesion. Although
glomeruli were normal by light microscopy, immunohistochemical studies demonstrated confluent finely granular epimembranous deposits of IgG and C3. This distribution was confirmed at the ultrastructural level with immunoelectron
microscopy. High resolution elemental analysis of electron
dense inclusions in tubular epithelial phagolysosomes demonstrated energy dispersion spectra characteristic of coexisting
mercury and selenium. Eluates from the biopsy material were
not immunoreactive against normal rat or human kidney. There
was no immunoreactivity of epimembranous deposits with antibodies having renal tubular epithelial antigen or urinary uromucoid specificity.
These observations suggest that a distinctive immunopathologic lesion is associated with mercury-associated membranous glomerulonephritis, that the role of the metal itself may
only be coincidental, and that the involved antigen remains unknown. Prednisone therapy had no documented persistent beneficial influence upon the level of proteinuria in one patient who
has been lost to follow-up. In one patient not treated with steroid therapy, withdrawal of exposure to the metal resulted in
disappearance of mercury from body fluids and clinical remission. (Key words: Membranous glomerulonephritis; Mercuric
nephropathy; Heavy metal toxicity; Tubular epithelial antigens)
Departments of Pathology, Immunopathology and
Hypertension/Nephrology, The Cleveland Clinic
Foundation, Cleveland, Ohio
thyroglobulin, carcinoembryonic antigen, and specific
tumor antigens had been identified as part of an immune
complex present in the glomerular basement membrane
(GBM). 19 However, the pathogenesis remains obscure
in most instances.
Nephrotoxicity has been associated with several
heavy metals including mercury, gold, bismuth, cadmium, and lead. Mercuric nephropathy is manifested
as one of two disease states based upon the severity of
initial exposure to the toxin. Large doses given intravenously produce selective necrosis of proximal tubules
in experimental animals. 614 Smaller doses of mercuric
compounds administered subcutaneously over a long
period of time result in the formation of an immune
complex nephropathy manifested either as mesangiopathic proliferative glomerulonephritis or a MGN. 1 1 0 1 3 1 8
MGN has been shown to occur in some patients receiving mercuric compounds as therapeutic agents.3,7
However, complete evaluation of renal biopsy material
has been performed in only one instance.7 To our knowledge, no previous report has described in direct analytic
fashion the presence of mercury in human renal biopsies.
This study describes results of a multidisciplinary
evaluation of material from two patients with MGN
who had industrial exposure to mercury.
THE MAJORITY OF CASES of membranous glomerulonephritis (MGN) are idiopathic. Occasionally,
specific antigens such as treponema, plasmodium, DNA,
Report of Two Cases
Case I
Received May 8, 1981; received revised manuscript and accepted
for publication July 10, 1981.
Address reprint requests to Dr. Tubbs: Department of Pathology,
The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland,
Ohio 44106.
A 33-year-old man worked in a chemical plant where he was exposed to mercury. He had been regularly checked for years for urine
mercury and urine protein elevations, and these studies had been normal. In January 1977 proteinuria was first noted. Two months later
peripheral edema, lethargy, and fatigue associated with mild arthralgias occurred. There was no history of recent upper respiratory in-
0002-9173/82/0400/0409 $00.75 © American Society of Clinical Pathologists
409
TUBBS ET AL.
410
fection or hypertension. A urinalysis demonstrated 0-1 erythrocytes
per high power field, occasional oval fat bodies, occasional fatty casts
and hyaline casts. Antinuclear antibody and LE cells were not detected. The C3, C4, total hemolytic complement blood urea nitrogen
and serum creatinine were within normal limits. The 24-hour urine
specimen contained 28.7 g of protein. A urine lead determination was
normal. The initial urine mercury determination was 547.58 /ig in a
24 hour urine specimen (normal 0-20). Renal venograms were normal.
Following the renal biopsy, the patient was treated with prednisone
with poor compliance, had persistent proteinuria when reevaluated 20
months after biopsy, but was subsequently lost to follow-up.
Case 2
A 24-year-old man was found to have proteinuria on a routine
screening program at a chemical plant where he operated an electrolysis unit. He was required to use a vapor hood, special protective
skin cream, gloves, and clothing, which were worn only in the work
area. Previous medical history had been negative except for a single
episode two years before of gross hematuria, following 75 push-iips,
which never recurred. An intravenous pyelogram was normal. The
urine sediment contained rare erythrocytes, leukocytes, and a rare
leukocyte cast. Occasional oval fat bodies were present. The 24-hour
urine specimen contained 3.13 g protein. Normal results were obtained
for rheumatoid factor, lupus erythematosus preparation, antinuclear
antibody, beta-2-microglobulin, total hemolytic complement, antiDNA, and C3. The urine mercury was 174 ^g in a 24 hour urine
specimen (normal 0-20). The blood urea nitrogen and-creatinine were
normal. The creatinine clearance was 116 ml per minute per 1.73
square meter.
The patient was advised to avoid exposure to mercury. Prednisone
was not administered. Fifteen months after onset of proteinuria the
urine mercury was within normal limits, BUN and serum creatinine
remained normal, but moderate proteinuria persisted (2.3 g/24
hours). When last evaluated 21 months after biopsy the patient was
clinically well, 24 hour urine protein was 0.110 g, and the serum
creatinine 0.9 mg/dl.
Materials and Methods
Renal tissue was obtained by percutaneous needle
biopsy. The submitted specimens were divided into
three portions. The first portion was fixed in zinc substituted Zenker's solution, paraffin embedded, and
stained with hematoxylin and eosin, Jones methenamine
silver, periodic acid Schiff, and Masson trichrome. The
second portion was frozen in liquid nitrogen and stored
at —70°C until sectioning. The third portion was diced
into 1 mm cubes and fixed in buffered glutaraldehyde,
postfixed in osmium tetroxide, dehydrated in graded
alcohols, and embedded in Spurr. In case two, an additional tissue alliquot was homogenated and extracted
with 6 molar sodium chloride. Urine samples and the
tissue homogenate mercury determinations were done
by fiameless atomic absorption.
Direct immunofluorescence and direct immunoperoxidase studies were done as previously described.15
Specimens immunostained with the direct immunoperoxidase procedure were counterstained with periodic
acid Schiff, dehydrated in graded alcohols, and mounted
in Permount as a permanent preparation. 15 Immunoelectron microscopic studies (IEM) were performed on
fresh frozen kidney tissue. Forty-micra cryostat sections
of fresh, snap-frozen tissue were collected on phos-
A.J.C.P. • April 1982
phate-buffered saline (PBS) and rapidly transferred
into PBS-buffered peroxidase conjugated F(ab')2 fragment rabbit anti-human IgG (Dako). After incubation
of 15 min, immunoreacted, unfixed sections were briefly
washed in PBS and fixed for 1 min in PBS buffered
2%, paraformaldehyde followed by a PBS-holding
wash. Sections were subsequently reacted for 10 min
in Tris-buffered 3,3-diaminobenzidine pH 7.6 with
0.003 per cent H 2 0 2 . Immunostained sections were then
postfixed in cacodylate-buffered 1% osmium tetroxide,
dehydrated in graded ethanols and Spurr embedded.
Thin sections for IEM were viewed and photographed
unstained.
Antigen identification and elution studies of the renal
biopsy specimens were done using modifications of techniques previously described. 219 IgG was eluted with 0.02
M citrate buffer pH 3.2 for four hr. The eluate was
dialyzed for 12 hr against PBS and applied to sections
of normal human kidney and rat kidney for 15 min,
washed briefly in PBS, and fluorescein isothiocyanate
conjugated anti-human immunoglobulins (Behring)
overlaid on the tissue for 10 min. Sections were subsequently washed and mounted in 90% glycerol in PBS
and examined with a Leitz fluorescent photomicroscope.
In situ demonstration of the antigen in epimembranous
immune complexes was attempted by indirect immunofluorescence with the use of rabbit antihuman renal
epithelial cell antibody (anti-Fx 1A serum), and urinary
uramucoid (Cappel), followed by fluorescein labeled
sheep antirabbit IgG antibody (Fc specific, Cappel).
For high resolution elemental analysis, thin sections
were ultramicrotomed from Spurr embedded tissue,
mounted on nylon grids, placed in a beryllium specimen
holder, and viewed unstained using a Philips EM 400T
scanning transmission electron microscope, equipped
with x-ray (EDAX) dispersion analysis. Elemental
analysis was done on electron-dense material present
in phagolysosomes in tubular epithelial cells. Adjacent
and remote tissue organelles and subepithelial electron
dense deposits were also analyzed as internal controls.
Results
Sections submitted for light microscopy included an
average of 15 glomeruli per section. Light microscopy
including assessment with silver impregnation stains
failed to reveal an abnormality. Immunoperoxidase and
immunofluorescence revealed diffuse generalized finely
granular 17 deposition of IgG and C3 along the glomerular basement membrane (Fig. 1). The immunoperoxidase sections counterstained with periodic acid Schiff
demonstrated an epimembranous distribution of most
of the deposits. IEM demonstrated coarsely granular
osmophilic color reaction product within subepithelial
deposits (Fig. 2).
Four glomeruli were photographed in each case by
t
FIG. 1 (upper, left). Immunofluorescence photomicrograph,
Case 2. Generalized finely granular deposits of C3 are present,
outlining the glomerular basement membrane. X400.
FIG. 2 (upper, right). Immunoelectron micrograph, direct immunoperoxidase technique, Case 2. Subepithelial granular peroxidative reaction product with peroxidase conjugated anti-IgG
antisera (arrows) corresponding to distribution of electron dense
subepithelial deposits observed in Fig. 4b. unstained, X23,400.
FIG. 3 (center, left). Electron micrograph, proximal tubular
epithelial cell cytoplasm. Finely granular electron-dense material
is present within a phagolysosome. Lead citrate and uranyl acetate, X27,700.
FIG. 4. Electron micrographs, Cases l(4a), (center, right) and
2(4b) (lower, left) Subepithelial electron-dense deposits (arrows)
overlie the glomerular basement membrane. Lead citrate and
uranyl acetate; 4a, X6500; 4b, X11,000
1
412
TUBBS ET AL.
A.J.C.P. • April 1982
analysis performed on the subepithelial deposits in the
glomeruli, red blood cell membranes, cytoplasmic organelles and electron dense inclusions within phagolysosomes in the tubular epithelium demonstrated typical
x-ray dispersion spectra of mercury only in tubular phagolysosomes (Fig. 5) and not within the other structures
analyzed. Also detected only in phagolysosomes containing mercury were levels of selenium that were not
present in other portions of the nephron (Fig. 5).
Attempts to characterize antibody specificity of
eluates revealed no immunoreactivity against either
normal rat or human kidney tissue. Sections stained for
anti-Fx 1A revealed staining of tubular epithelium, predominately in the brush border, but no staining of the
GBM or any component of glomeruli.
Discussion
FIG. 5. TEM/EDAX analysis of tubular epithelial electron dense
inclusion. Inset illustrates finely granular inclusions in clustered phagolysosomes of tubular epithelium. Unstained, XI 7,000; energy dispersion spectra characteristic of mercury and selenium are identified
for inclusions. Indicated are the M a & 0 (2.199 KEV, 2.287 KEV),
L a (9.988 KEV), L 0 (11.821 KEV), and L y (13.829 KEV) energy
levels for mercury in addition to the Ka (11.220 KEV) and L a (1.381
KEV) energy levels for selenium. Also present is the spectrum characteristic of osmium (not indicated).
transmission electron microscopy. Vessels and interstitium were unremarkable. Within the proximal tubular
epithelial cell cytoplasm were heterogeneous electron
dense phagolysosomes visible on unstained sections.
These organelles were spherical or irregular in contour
and contained a finely granular electron opaque material (Fig. 3). The glomerular basement membranes
were of normal thickness, but foot processes were broadened. Confluent subepithelial electrondense deposits
were present along the peripheral capillary loop basement membranes (Fig. 4). The appearance of the subepithelial deposits was homogenous without a definable
substructure. No subendothelial or mesangial deposits
were identified. The mesangial cells and matrix were
normal by ultrastructural assessment. Tubular arrays
were not present in endothelial cell cytoplasm. EDAX
Earlier reports, which included only descriptions of
light microscopy in patients with mercury associated
nephrotic syndrome, have described no glomerular histopathologic abnormalities. 20 The clinical settings in
which the lesion occurs have included administration
of teething powders containing the metal to pediatric
patients, skin-lightening creams used by African blacks,
and ammoniated mercury ointments for psoriasis.3,7,20
The present report describes the lesion occurring in an
occupational setting. Results of elution and antigen
identification studies were not successful in identifying
specific offending agents. The exclusion of mercury
from environmental exposure in one patient was resulted in disappearance of the metal from body fluids
and clinical remission.
Four disease mechanisms have been proposed to account for the immunopathologic findings in mercuric
nephropathy. 713 The association of MGN and mercury
exposure may be coincidental. Mercury accumulates in
renal tubular epithelial lysosomes of proteinuric rats
regardless of the etiology of proteinuria. 9 Secondly, the
metal may combine with a serum protein and antibodies
formed against the hapten. Inability to identify mercury
in the glomerular subepithelial deposits in the present
study may indicate cross reactivity of the offending
antibody with the mercury hapten and serum proteins,
the latter being in excess and more readily combined
with antibody and trapped in the GBM. Alternatively,
antirenal tubular epithelial cell or antinuclear antibody
formation may be involved, 7 ' 218 although the results
of our study do not support these hypotheses. The fourth
explanation is based upon an experimental model developed by Roman-Franco and associates.13 Mercury
induced antibodies to basement membrane and extracellular collagen matrix, and a M G N develops as im-
MERCURY AND GLOMERULONEPHRITIS
Vol. 77 • No. 4
mune complexes are formed locally in the GBM or
epimembranous space.13 The present study did not confirm or conclusively exclude this pathogenesis.
Characteristic filamentous electron dense inclusions
have been identified in renal tubular epithelium of patients with gold nephropathy and have been shown to
contain gold by elemental analysis.16 The present study
describes mercury in electron dense inclusions in phagolysosomes in the same location, but with a more finely
granular substructure. Our finding of selenium in phagolysosomes also containing mercury may indicate a
natural affinity of these elements in biological systems.
Other reports have noted the natural and experimental
accumulation of mercury and selenium associated with
the amelioration of the toxic effects of mercury poisoning.4,5,8" In our study, the detection of mercury and
selenium may have been facilitated by terminal condensation occurring within the tubular phagolysosomes.
Thus, it is not possible to exclude the presence of mercury and/or selenium in other compartments of the
nephron since the detection of these elements in other
cellular regions may be beyond present levels of sensitivity. Analytical correlations are presently being performed to determine the levels of selenium found in
patients having mercury and other metal associated
nephropathies. Our experiences with nephrotic syndrome associated with metal contaminants suggest it
prudent to examine unstained material by electron microscopy in each case of idiopathic MGN and to evaluate biopsy material with elemental analysis in selected
cases.
Acknowledgement.
serum.
Dr. Richard Zager kindly supplied Anti-Fx 1A
References
1. Bariety J, Druet P, Laliberte F, Sapin C: Glomerulonephritis with
gamma and beta-l-C globulin deposits induced in rats by mercuric chloride. Am J Pathol 65:293-300, 1971
2. Bartolotti SR: Quantitative elution studies in experimental immune complex and nephrotoxic nephritis. J Clin Exp Immunol
29:334-341, 1977
3. Becker CG, Becker EL, Maher JF, Schreiner GE: Nephrotic
syndrome after contact with mercury. A report of five cases,
three after the use of ammoniated mercury. Arch Intern Med
110:178-186, 1962
413
4. Carmichael NG, Fowler BA: Effects of separate and combined
chronic mercuric chloride and sodium selenate administration
in rats: histological, ultrastructural and x-ray microanalytical
studies of liver and kidney. J Environ Pathol Toxicol 3:399412, 1979
5. Chang LW, Dudley AW, Dudley MA, Ganther HE, Sunde ML:
Modification of the neurotoxic effects of methyl mercury by
selenium. Neurotoxicol, edited by L Roizin, H Shiraki, N
Grcevic. New York, Raven Press, 1977, pp 137-145
6. Ganote CE, Reimer KA, Jennings RB: Acute mercuric chloride
nephrotoxicity. An electron microscopic and metabolic study.
Lab Invest 31:633-647, 1975
7. Kibukamusoke JW, Davies DR, Hutt MSR: Membranous nephropathy due to skin-lightening cream. Br Med J 2:646-647,
1974
8. Kosta L, Byrne AR, Zelenko V: Correlation between selenium
and mercury in man following exposure to inorganic mercury.
Nature 254:238-239, 1975
9. Madsen KM: Mercury accumulation in kidney lysosomes of proteinuric rats. Kid Intern 18:445-453, 1980
10. Makker SP, Aikawa M: Mesangial glomerulonephropathy with
deposition of IgG, IgM, and C3 induced by mercuric chloride.
A new model. Lab Invest 41:45-50, 1979
11. Megos L, Webb M: The interactions of selenium with cadmium
and mercury. CRC Crit Rev Toxicol 8:1-42, 1980
12. Naruse T, Miyakawa Y, Kitamura K, Shibata S: Membranous
glomerulonephritis mediated by renal tubular epithelial antigen-antibody complex. J Allergy Clin Immunol 54:311-318,
1974
13. Roman-Franco AA, Twirello M, Albini B, Ossi E, et al: Antibasement membrane antibodies and antigen-antibody complexes in rabbits injected with mercuric chloride. Clin Immunol
Immunopathol 9:404-481, 1978
14. Siegel FL, Bulger RE: Scanning and transmission electron microscopy of mercuric chloride induced acute tubular necrosis
of rat kidneys. Virchows Arch B Cell Path 18:243-262, 1975
15. Tubbs RR, Gephardt GN, Valenzuela R, Deodhar SD: An approach to diagnosis of renal disease with the immunoperoxidase
and periodic acid Schiff counterstain (IMPAS stain). Am J
Clin Pathol 73:240-244, 1980
16. Tubbs RR, Valenzuela R, McCormack LJ, Pohl MA, Barenberg
S: Gold nephropathy. N Engl J Med 296:1413-1414, 1977
17. Valenzuela R, Gogate P, Deodhar SD, Pohl MA, Yeip M: Diffuse
generalized pseudolinear (finely granular) deposition of immunoglobulins and complement in the glomerular basement
membrane: clinicopathologic significance. (Abstract) Am J
Clin Pathol 72:651, 1979
18. Weening JJ, Fleuren GJ, Hoedemaeker PJ: Demonstration of
anti-nuclear antibodies in mercuric chloride-induced glomerulopathy in the rat. Lab Invest 39:405-411, 1978
19. Woodroffe A J, Wilson CB: An evaluation of elution techniques
in the study of immune complex glomerulonephritis. J Immunol
118:1788-1794, 1977
20. Worthen HG, Vernier RL. Goode RA: Infantile nephrosis. Clinical, biochemical, and morphologic studies of the syndrome.
Am J Dis Child 98:731-748, 1959