Immunocytochemical Enhancement of Basement Membrane
Antigens by Pepsin: Applications in Diagnostic Pathology
SANFORD H. BARSKY, M.D., N. C. RAO, PH.D., CARLOS RESTREPO, M.D.,
AND LANCE A. LIOTTA, M.D.
Basement membrane zone antigens, laminin, Type IV collagen,
and Type V collagen each are enhanced selectively in formalinfixed, paraffin-embedded sections by pretreatment with pepsin
(1 mg pepsin/mL 0.5 M acetic acid). Highly specific antibodies
to laminin, Type IV collagen, and Type V collagen each demonstrate intense linear staining of epithelial and endothelial
basement membranes in fresh tissue, but this staining is diminished markedly and often completely abolished in paraffinembedded tissue. Pretreatment of this paraffin-embedded tissue
with standard immunoperoxidase enhancement technics such as
trypsin fails to increase immunoreactivity of these antigens. Pretreatment with other proteases with known collagenolytic activities such as bacterial collagenase and Type IV collagenase
is likewise ineffective. In contrast, pepsin pretreatment markedly
enhances immunoreactivity of laminin, Type IV collagen, and
Type V collagen. Previous immunocytochemical studies of the
basement membrane useful in the distinction of invasive tumors
from their benign counterparts, for example, tubular carcinoma
of the breast from sclerosing adenosis, which only could be done
prospectively in fresh tissue now can be performed retrospectively
in paraffin-embedded tissue. (Key words: Type IV collagen; pepsin; basement membranes) Am J Clin Pathol 1984; 82
191-194
BASEMENT MEMBRANES are ubiquitous extracellular
matrices that separate organ parenchymal cells from interstitial stroma and are synthesized by a number of cells
including epithelial, endothelial, and certain mesenchymal
cells.12 Although the molecular composition of basement
membranes differ depending on the cell of origin, all
basement membranes share common antigenic determinants that include laminin and Type IV collagen. 7 "
Type V collagen, while not residing in the basement
membrane proper, lies in between that structure and the
adjacent interstitium and is considered a common basement membrane zone antigen.4
Determining the status of the basement membrane in
tissue sections is useful in both normal and disease states.
Blood vessel capillaries, for example, can be identified in
tissue sections by utilizing antibodies to Type IV collagen
and laminin via immunoperoxidase staining and further
distinguished from lymphatic capillaries and artefactual
Laboratory of Pathology, NCI, National Institutes of Health,
Bethesda, Maryland
tissue spaces by the lack of this immunoreactivity in these
latter structures.' Similarly, invasive epithelial tumors can
be distinguished from benign and preinvasive lesions in
a wide variety of tissues by the consistent loss of a surrounding basement membrane in the invasive tumors as
evidenced by absent immunoreactivity to Type IV collagen and laminin. Benign and preinvasive lesions, on
the other hand, consistently exhibit intact, linear Type
IV collagen and laminin basement membrane immunoreactivity.3
The immunocytochemical demonstration of basement
membrane immunoreactivity, like most immunocytochemical staining, works best in fresh or fresh-frozen tissue. Although Type IV collagen, laminin, and Type V
collagen antigens are still reasonably preserved and demonstrable in formalin-fixed tissue, paraffin embedding
results in marked loss and often complete abolishment
of immunoreactivity of all three antigens.
Since we wanted to extend our previous observations
of the basement membrane in invasive versus noninvasive
tumors retrospectively to tissues embedded in paraffin to
aid in the distinction of such entities as sclerosing adenosis
of the breast from tubular carcinoma, we attempted to
enhance immunocytochemical staining of these basement
membrane zone antigens. We pretreated paraffin-embedded sections with different proteases over a wide concentration range in an attempt to enhance basement
membrane immunoreactivity. We used a standard immunocytochemical enhancement technic with trypsin; in
addition we used other proteases having known collagenolytic activities. These included bacterial collagenase,
Type IV collagenase, and pepsin.
Materials and Methods
Received September 26, 1983; received revised manuscript and accepted for publication December 12, 1983.
Address reprint requests to Dr. Barsky: Laboratory of Pathology, NCI,
Building 10, Room 2N-212, National Institutes of Health, Bethesda,
Maryland 20205.
Formalin-fixed, paraffin-embedded tissues containing
abundant basement membranes (human amnion [three
cases] and sclerosing adenosis of the breast [five cases])
and tissues devoid of such structures (tubular carcinoma
of the breast [5 cases]) were prepared for immunoper-
191
192
BARSKY ET AL.
A.J.C.P. -August 1984
FIG. 1 (upper). A (left). Type IV collagen immunoreactivity is completely absent in paraffin-embedded sections of human amnion. No linear
basement membrane staining is apparent underneath the epithelial layer. Anti-type IV collagen, immunoperoxidase (X250). B (right). In contrast
to A, pretreatment of the section with pepsin restores Type IV collagen immunoreactivity as evidenced by linear basement membrane immunoreactivity
underneath the epithelial layer. Laminin and Type V collagen immunoreactivity could be restored in a similar manner. Other proteases including
bacterial collagenase, Type IV collagenase, and trypsin were ineffective in restoring basement membrane immunoreactivity. Anti-type IV collagen,
immunoperoxidase (X250).
FlG. 2 (lower). A (left). Pepsin pretreatment of paraffin-embedded tissue followed by antibasement membrane staining is useful in distinguishing
invasive tumors from their "look alike" benign counterparts. Sclerosing adenosis of the breast displays intense linear basement membrane immunoreactivity surrounding each and every acinus. Anti-laminin, immunoperoxidase (X500). B (right). In contrast, tubular carcinoma of the breast
fails to display such immunoreactivity. Anti-laminin, immunoperoxidase (X500).
oxidase staining according to a standard protocol. The
paraffin was dissolved with xylene for 15 minutes, two
changes, followed by hydration of the tissue in a series
of graded alcohols ranging from 100% to 50%. The tissue
then was rinsed in phosphate-buffered saline (PBS), pH
7.4. Endogenous peroxidase activity was abolished by 30-
BRIEF SCIENTIFIC REPORTS
Vol. 82 • No. 2
minute incubation in 1 % H 2 0 2 in methanol. Nonspecific
staining was abolished by 30-minute incubation in 2%
swine serum in PBS.
The tissue then was treated with either trypsin (0.010.1%), bacterial collagenase (0.01-1 mg/mL), purified
Type IV collagenase (0.01-1 mg/mL), pepsin (1 mg/mL,
0.5 acetic acid), 0.5 M acetic acid alone, or PBS. Trypsin
was obtained from Flow Laboratories, McClean, Virginia,
bacterial collagenase and . pepsin were obtained from
Sigma Chemical Company, St. Louis, Missouri; Type IV
collagenase was purified and characterized in a previous
study.8 The incubations were carried out at room temperature and at 37 °C for varying time courses ranging
from 30 minutes to 24 hours.
Antibodies to Type IV collagen, laminin, and Type V
collagen characterized and purified in previous studies 23
were applied separately in tandem to sections of the previously mentioned tissues. The first antibody was applied
for one hour (1/20 dilution in PBS) at room temperature
in a humidified chamber. The sections then were washed
in PBS and a goat antirabbit antibody (1/20 dilution)
(heavy and light chain specific, Cappel Laboratories,
Cochranville, PA) was applied for 30 minutes. After thorough washing with PBS, rabbit-peroxidase-antiperoxidase complex (PAP)10 was applied (1/100 dilution) for
30 minutes. After thorough washing, the color was developed with 3,3'diaminobenzidine:H 2 0 2 (Sigma Chemical Company). The slides were counterstained with hematoxylin, dehydrated in a series of graded alcohols and
xylene, and Permount coverslipped. The staining patterns
of the basement membrane with each of the three antibodies were evaluated separately in the groups pretreated
with the different enzymes.
Results
There was a dramatic increase in laminin, Type IV
collagen, and Type V collagen immunoreactivity in the
sections pretreated with pepsin. No appreciable immunoreactivity to any of these antigens could be detected in
the sections pretreated with the other proteases in any
concentration, nor did any significant staining occur in
the slides pretreated with acetic acid alone or in the slides
without pretreatment (Fig. 1) (Table 1). The optimal immunoreactivity occurred with two hours pretreatment
with pepsin at 37°C.
In the sections treated with the higher concentrations
of trypsin and bacterial collagenase for the longer incubation periods (>four hours), the tissue was digested partially off the slide. The portion of the tissue remaining
nevertheless showed no basement membrane immunoreactivity. In contrast, the sections treated with pepsin
never experienced such tissue digestion. The three basement membrane zone antigens each were enhanced with
193
Table 1. Immunoreactivity of Basement Membrane
Zone Antigens in Paraffin with Treatment
with Different Proteases
Anti-laminin
Anti-type IV
Anti-type V
++++
—
+++
—
+++
—
—
—
—
—
—
—
—
—
—
—
—
—
Pepsin (1 mg/ml 0.5
M acetic acid)
Trypsin (0.01-1%)
Type IV collagenase
(0.01-1 mg/mL)
Bacterial collagenase
(0.01-1 mg/mL)
Acetic acid alone
(control)
PBS (control)
Positive staining: + to +++++; absent staining —.
this enzyme, but the laminin enhancement was slightly
greater than that for Type IV and Type V collagen. The
immunoreactivity of these basement membrane zone antigens was still the strongest in fresh tissue, but the enhancement of immunoreactivity of paraffin-embedded
tissue with pepsin produced reasonably intense staining
exceeding that seen in formalin-fixed nonparaffinembedded tissue that was not pepsin pretreated (Table
2). Pretreatment with pepsin then is a method that significantly reverses the quenching effect of paraffin and
formalin and restores immunoreactivity to the group of
basement membrane zone antigens: laminin, Type IV
collagen, and Type V collagen.
When these findings were applied to paraffin-embedded
sections of five cases of tubular carcinoma of the breast
and five cases of sclerosing adenosis, lesions that commonly resemble each other in routine histologic sections,
a clear-cut distinction was apparent. Tubular carcinoma
displayed complete absence of basement membrane immunoreactivity, whereas sclerosing adenosis demonstrated
intact, linear, basement membrane immunoreactivity
around every acinar structure in all five cases (Fig. 2).
Discussion
Pretreatment with pepsin is a method that markedly
enhances basement membrane immunoreactivity in parTable 2. A Comparison of Basement Membrane
Immunoreactivity in Fresh, Formalin-Fixed,
and Paraffin-Embedded Tissue with
and without Pepsin Treatment
Anti-laminin Anti-type IV Anti-type V
Fresh tissue
Formalin-fixed tissue
Paraffin-embedded tissue
With pepsin
Without pepsin
+++++
++
+++++
++
+++++
++
++++
—
+++
—
+++
—
Positive staining: + to +++++; absent staining —.
194
A.J.C.P. • August 1984
BARSKY ET AL.
affin-embedded tissue and permits an accurate assessment
of the status of the basement membrane in stored material.
For example, the determination of whether a nest of epithelial cells exhibits true invasion or pseudoinvasion, a
determination that rests on the absence or presence of
basement membrane components, 3 now can be made in
paraffin-embedded material.
In the breast, for example, sclerosing adenosis with
linear basement membrane immunoreactivity demonstrated with antibodies to laminin, Type IV collagen, and
Type V collagen can be distinguished clearly from tubular
carcinoma (Fig. 2). There are many other situations where
the accurate evaluation of basement membranes in paraffin tissue is diagnostically important, for example, in
the determination of vascular invasion by tumor.
The selective immunoenhancement effect of pepsin on
laminin alone has been noted by others,5 but we have
extended these findings to Type IV collagen and Type V
collagen allowing us to conclude that pepsin has an important immunoenhancement action on the basement
membrane region in general. Furthermore, by comparing
the effect of pepsin to that of other proteases, especially
trypsin, a commonly used enzyme in immunocytochemistry, we have observed the selective role of pepsin on
immunoenhancement of the basement membrane.
The exact mechanism by which pepsin selectively enhances basement membrane antigens is not known. All
the proteases tested have known basement membrane
collagenolytic activities, but each attacks Type IV collagen
at different sites.9 Of all these proteases, pepsin is the
most effective in solubilizing Type IV collagen from basement membrane 6 and perhaps its action in tissue is based
on this ability to solubilize portions of the molecule,
thereby exposing previously masked antigenic sites. It is
a bit surprising that trypsin, a well-known and generally
accepted all purpose immunoenhancer does not enhance
the basement membrane zone antigens. Trypsin is not
very effective, however, in solubilizing Type IV collagen.
Type IV collagenase and bacterial collagenase, although
effective in solubilizing Type IV collagen, have extremely
high collagenolytic activities,8 which may destroy critical
antigenic sites that our antibodies first must recognize in
order to demonstrate immunoreactivity. These latter enzymes may in fact be too potent even at lower concentrations.
It is important to realize that this enhancement of basement membrane immunoreactivity by pepsin in paraffinembedded tissue still does not produce as intense staining
as is demonstrated directly in fresh tissue. This observation
represents the rule in immunocytochemistry that "fresh"
antigens are better than "fixed" "antigens." Nevertheless,
our ability to recover significant immunoreactivity of these
basement membrane antigens, laminin, Type IV collagen,
and Type V collagen gives us an important tool to resurrect information about the basement membrane in
many diseases buried for years in paraffin.
References
1. Barsky SH, Baker A, Siegal GP, Togo S, Liotta LA: Use of antibasement membrane antibodies to distinguish blood vessel capillaries from lymphatic capillaries. Am J Surg Pathol 1983; 7:667677
2. Barsky SH, Rao CN, Grotendorst GR, Liotta LA: Increased content
of type V collagen in desmoplasia of human breast carcinoma.
Am J Pathol 1982; 108:276-283
3. Barsky SH, Siegal G, Jannotta F, Liotta LA: Loss of basement
membrane components by invasive tumors but not by their
benign counterparts. Lab Invest 1983; 49:140-148
4. Bornstein P: Structurally distinct collagen types. Annu Rev Biochem
1980; 49:957-1003
5. Ekblom P, Miettinen M, Rapola J, Foidart JM: Demonstration of
laminin, a basement membrane glycoprotein, in routinely processed formalin-fixed human tissues. Histochemistry 1982;
75:301-307
6. Kefalides NA: Chemistry of basement membranes: Structure and
biosynthesis. Adv Microcirc 1980; 9:295-322
7. Kefalides NA, Denduchis B: Structural components of epithelial
and endothelial basement membranes. Biochemistry 1969;
8:4613-4621
8. Liotta LA, Tryggvason K, Garbisa S, Gehron-Rohey P, Abe S:
Partial purification and characterization of a neutral protease
which cleaves type IV collagen. Biochemistry 1981; 20:100-104
9. Minor RR: Collagen metabolism. Am J Pathol 1980; 98:227-280
10. Stemberger LA: The unlabelled antibody peroxidase-antiperoxidase
(PAP) method, Immunocytochemistry. Edited by LA Stemberger.
New York, John Wiley and Sons, 1979, pp 297-307
11. Timpl R, Rohde H, Gehron Rohey P, Rennard SI, Foidart JM,
Martin GR: Laminin—a glycoprotein of basement membranes.
J Biol Chem 1979; 254:9933-9937
12. Vracko R: Basal lamina scaffold-anatomy and significance for
maintenance of orderly tissue structure. Am J Pathol 1974;
77:314-346