OUP user menu

Immunofluorescence With Dual Microwave Retrieval of Paraffin-Embedded Sections in the Assessment of Human Renal Biopsy Specimens

Suozhu Shi MD, Qingli Cheng PhDMD, Ping Zhang PhDMD, Nan Wang MD, Ying Zheng MD, Xue-Yuan Bai PhDMD, Xiangmei Chen PhDMD
DOI: http://dx.doi.org/10.1309/AJCPRZG8EXN7BAID 71-78 First published online: 1 January 2013

Abstract

Immunofluorescence of frozen tissue sections (IF-F) is a classic technique for renal immunopathologic examination. However, it has certain disadvantages, such as diffuse antigen distribution and few or even no glomeruli in the section. We developed a new technique of immunofluorescence staining using dual microwave retrieval in paraffin-embedded renal tissue sections (IF-DMP) and compared IF-DMP with IF-F in 406 renal biopsy samples. IF-DMP detected significantly more glomeruli than did IF-F (P< .001). There was no significant difference for the specificity and sensitivity in the detection of immunoglobulins, complements, κ, and λ between IF-F and IF-DMP. Concordant observations were 98% for all immunofluorescence, complements, κ, and λ staining and 100% for immunoglobulin staining. Both techniques were completely accurate in confirming diagnoses of various glomerular diseases. IF-DMP provided clearer images of tissue structure and more precise localization of antigens, and it is a suitable alternative for traditional IF-F in clinical renal immunopathologic diagnosis.

Key Words
  • Immunofluorescence staining
  • Frozen tissue section
  • Paraffin-embedded section
  • Microwave retrieval
  • Renal biopsy specimen

Direct immunofluorescence staining using frozen tissue sections (IF-F) has long been the gold standard for detection of immune complexes and complements in renal immuno-pathologic diagnosis.1 However, there are some disadvantages of IF-F in clinical practice. First, the IF-F section may be too thick for detailed analysis owing to tissue structure overlap. Second, the antigen may appear to have a diffuse distribution, making determination of immunoglobulin distribution difficult. Third, a limited amount of tissue is sampled during biopsy, and only a few or even no glomeruli may be provided, decreasing the accuracy of evaluation. Last, the frozen sections cannot be stored for subsequent retrospective studies. In recent years, some researchers have used trypsin or pronase to digest paraffin tissue sections for antigen retrieval. However, enzyme digestion may damage the tissue structure and affect the accuracy of diagnosis.25

In the past decade, we have used IF-F for renal immuno-pathologic diagnosis in 10,356 patients with kidney disease. We also have developed a new technique of dual microwave retrieval in formaldehyde-fixed and paraffin-embedded renal tissues (IF-DMP) for immunofluorescence staining and applied it in clinical immunopathologic diagnosis. Compared with IF-F, the IF-DMP sections are very thin, allow more precise localization of antigens, and are less limited by the small amounts of available tissue. In addition, IF-DMP has favorable sensitivity, specificity, and reproducibility in the detection of immunoglobulins and complements in renal tissues.

The purpose of this study was to further compare IF-DMP and IF-F in renal pathologic examinations. We used immunofluorescence staining with IF-DMP and IF-F in the same samples from 406 patients with different renal diseases and compared their sensitivity, specificity, and reliability in renal immunopathologic diagnosis.

Materials and Methods

Collection of Renal Samples

Renal samples from 406 patients with various renal diseases were collected by biopsy in the Department of Nephrol-ogy, the Chinese PLA General Hospital, Beijing, China, from May 2011 to November 2011. There were 75 patients with membranous nephropathy (MN), 88 patients with lupus nephritis (LN), 125 patients with IgA nephropathy (IgAN), 32 patients with acute postinfectious glomerulonephritis (APGN), 30 patients with type 1 membranoproliferative glo-merulonephritis (MPGN), 35 patients with anti–glomerular basement membrane nephropathy (anti-GBM disease), and 21 patients with primary amyloidosis (PA).

Dual Microwave Retrieval and Immunofluorescence Staining With IF-DMP

Renal tissues from a biopsy specimen were fixed in 10% formaldehyde, embedded in paraffin, cut into 1- to 2-μm sections with a microtome, and attached to 0.01% poly-L-lysine– coated slides. After drying, deparaffinization, dehydration in gradient alcohol, and rinsing in phosphate-buffered saline (PBS; 0.01M, pH 7.2-7.4), the sections were subjected to antigen retrieval by immersion in an EDTA (0.01M, pH 8.0) antigen retrieval solution and heated in a microwave oven (SHARP, Osaka, Japan) at 320 W for 3 minutes and then at 850 W for 2.5 minutes. Then, the sections were cooled to 4°C for 10 minutes, washed in PBS, and fixed in 4% formaldehyde for 5 minutes again. After being washed in PBS, a second antigen retrieval was performed in a microwave oven as described above, and the sections were kept in EDTA buffer at 100°C for 10 minutes. Then, the sections were cooled to room temperature for 20 minutes, washed in PBS for 5 minutes, and treated with FITC-conjugated IgG (1:40), IgA (1:40), IgM (1:40), C3 (1:40), C1q (1:40), κ (1:50), and ℓ (1:50) rabbit polyclonal antibodies (DAKO, Glostrup, Denmark) at room temperature for 30 minutes. After washing in PBS twice (3 minutes each), sections were mounted with glycerin and observed under a fluorescence microscope (Nikon 80i; Nikon, Tokyo, Japan).

Processing of Frozen Tissue Sections and IF-F Staining

Fresh renal tissues from a biopsy specimen were embedded in Tissue-Tek O.C.T. Complex (Sakura Finetek Europe BV, Zoeterwoude, the Netherlands) and cut into 3- to 4-μm sections with a cryotome (CM3050S; Leica, Nussloch, Germany). After drying at room temperature for 20 minutes and washing in PBS for 5 minutes, sections were treated with FITC-conjugated IgG (1:40), IgA (1:40), IgM (1:40), C3 (1:40), C1q (1:40), κ (1:50), and λ (1:50) rabbit polyclonal antibodies (DAKO) at room temperature for 30 minutes. Following washing in PBS twice (3 minutes each), sections were mounted with glycerin and observed under a Nikon 80i fluorescence microscope.

Determination of Positive Staining Areas

The positive staining areas for the target antigens were measured with image analysis software (NIS-Elements, Nikon). The positive staining areas for IgG, IgA, IgM, C3, C1q, κ, and λ were measured in 5 complete glomeruli in each section, and the total area of each glomerulus was measured respectively. The ratio of the target antigen-positive staining area to the total glomeruli area was calculated as the relative content of immunoglobulin or complement.

Statistical Analysis

Data on the number of glomeruli and target antigen-positive staining area were expressed as means ± standard deviations. All statistical analyses were performed with SPSS version 17.0 (SPSS, Chicago, IL). The Student t test was used for comparisons. The λ2 test and McNemar test were used to compare differences in sensitivity and specificity in the detection of target proteins by IF-DMP and IF-F.

Results

Number of Glomeruli in the Sections From IF-DMP and IF-F

We detected 27.51 ± 6.98 glomeruli per section by IF-DMP and 6.92 ± 1.15 glomeruli per section by IF-F (P < .001) Table 1. Clearly, IF-DMP provided a significantly better overall count of glomeruli than did IF-F.

View this table:
Table 3
View this table:
Table 1

Positive Staining of IF-DMP and IF-F

Among biopsy samples from patients with 8 different renal diseases, IF-F and IF-DMP provided comparable levels of staining and localization of IgG, IgA, IgM, C3, C1q, k, and l positive regions. In the patients with LN, MN, IgAN, APGN, MPGN, anti-GBM, and PA, the location and staining intensities of positive regions were identical for IF-F and IF-DMP. There was no significant difference for the specificity and sensitivity in the detection of renal immunoglobulins, complements, k, and l between IF-F and IF-DMP Table 2. Concordant observations between IF-F and IF-DMP were 98% for all immunofluorescence staining (2,795 of 2,842 observations); 100% for IgG, IgA, and IgM (406 of 406 observations); 97% for C3 (395 of 406 observations); 96% for C1q (388 of 406 observations); 98% for k (398 of 406 observations); and 98% for l (396 of 406 observations).

Measure of Positive Staining Areas for IF-DMP and IF-F

We employed image analysis software (NIS-Elements) to measure the positive staining areas of IgG, IgA, IgM, C3, C1q, k, and l in the glomeruli. The results indicated that there were no significant differences in the relative content of immunoglobulins, complements, k, and l, as determined by IF-F and IF-DMP Table 3.

Comparison of Antigen Distribution in the Section of IF-DMP and IF-F

The tissue sections derived from IF-DMP were relatively thin, so the tissue structure was clearer and the antigens could be localized more precisely Image 1. In contrast, the sections derived from IF-F were thicker, so the tissue structure was less clear, and the antigens appeared to have more diffuse distributions Image 2.

View this table:
Table 2
Image 1

Representative renal tissue images from a light microscope with periodic acid-Schiff (PAS) staining and immunofluorescence staining using dual microwave retrieval of formaldehyde-fixed and paraffin-embedded renal tissues (IF-DMP) and immunofluorescence staining of frozen tissue sections (IF-F). A, PAS staining of a patient with membranous nephropathy. B, IF-DMP of the same patient, with IgG granules (green) at the capillary loops of the glomeruli. C, IF-F of the same patient, with IgG granules (green) at the capillary loops of the glomeruli. A and B, consecutive sections, ×400. D, PAS staining of a patient with IgA nephropathy. E, IF-DMP of the same patient, with patchy IgA granules (green) lateral to the mesangial region. F, IF-F of the same patient, with patchy IgA granules (green) lateral to the mesangial region. D and E, consecutive sections, ×400. G, PAS staining of a patient with lupus nephritis (type IV + V). H, IF-DMP of the same patient, with IgG granules (green) at the capillary loops of the glomeruli. I, IF-F of the same patient, with IgG granules (green) at the capillary loops of the glomeruli. G and H, consecutive sections, ×400. J, PAS staining of a patient with primary amyloidosis (amyloid protein, light chain derived), κ type. K, IF-DMP of the same patient, with κ depositing in the glomerular mesangial region. L, IF-F of the same patient, with κ depositing in the glomerular mesangial region. J and K, consecutive sections, ×400.

Image 2

Representative images from immunofluorescence staining using dual microwave retrieval of formaldehyde-fixed and paraffin-embedded renal tissues (IF-DMP) and immunofluorescence staining of frozen tissue sections (IF-F) for the same patient with membranous nephropathy. A, Periodic acid-Schiff (PAS) staining of a paraffin-embedded section. B, IF-DMP of a paraffin-embedded section, with IgG granules (green) regularly located at the capillary loops of the glomeruli. A and B, consecutive sections, ×800. C, PAS staining of a frozen section. D, IF-F of a frozen section, with IgG granules (green) irregularly located at the capillary loops of the glomeruli. C and D, consecutive sections, ×800.

Double-Staining Detection in IF-DMP and IF-F

Double staining can be performed in both IF-F and IF-DMP. However, double staining with IF-F could lead to false-positive results for antigen coexpression because of overlapping tissue structures. IF-DMP provides a clearer image of tissues, so localization of regions with antigen coexpression is more precise, and false-positive results are less likely Image 3. In addition, images provided by IF-DMP can be readily compared with those from conventional staining (eg, periodic acid-Schiff staining), indicating that IF-DMP provides more reliable detection of double fluorescence-conjugated antigens than IF-F.

Discussion

Immunopathologic staining of paraffin-embedded tissue sections often employs immunoperoxidase (IP), which uses horseradish peroxidase to label target antigens, but this may lead to false-positive results. In addition, the concordance between IP staining and IF-F is only 70% to 80%, and the sensitivity of IP staining is only 73% to 86%. IP staining cannot be used to identify linear deposits of IgG in anti-GBM disease.1 Thus, the application of IP staining in clinical diagnosis has certain limitations. Previous studies reported that formaldehyde-fixed and paraffin-embedded sections (IF-P) that had immunofluorescence staining were much weaker than frozen tissue sections or even undetectable for the deposition of C3.3,4 Formaldehyde-fixed and paraffin-embedded sections are traditionally considered unsuitable for immunofluorescence staining because calcium and other divalent ions form complexes with proteins during fixation in formaldehyde, and the complexes can block the antigenic determinants.68

Image 3

Representative images from immunofluorescence staining using dual microwave retrieval of formaldehyde-fixed and paraffin-embedded renal tissues (IF-DMP). A, Periodic acid-Schiff (PAS) staining of a patient with membranous nephropathy. B, IF-DMP of the same patient, with IgG granules (green) at the capillary loops of the glomeruli. C, Double staining and visualization by IF-DMP, with IgG granules (green) outside the glomerular basement membrane (GBM) and type IV collagen (Coll IV, red) at the GBM. A-C, consecutive sections from the same patient, ×400. D, PAS staining of a patient with IgA nephropathy. E, IF-DMP of the same patient, with IgA granules (green) depositing in the glomerular mesangial region. F, Double staining and visualization by IF-DMP, with patchy IgA granules (green) lateral to the mesangial region and Coll IV (red) at the GBM. Coexpression of IgA (green) and Coll IV (red) in the mesangial region resulted in yellow staining. D-F, consecutive sections from the same patient, ×400. G, PAS staining of a patient with membranous lupus nephritis. H, IF-DMP of the same patient, with C1q granules (green) outside the GBM. I, Double staining and visualization by IF-DMP, with C1q granules (green) outside the GBM and Coll IV (red) at the glomerular basement membrane. G-I, consecutive sections from the same patient, ×400. J, PAS staining of a patient with primary amyloidosis (amyloid protein, light chain derived), κ type. K, IF-DMP of the same patient, with κ (green) depositing in the glomerular mesangial region. L, Double-staining and visualization by IF-DMP, with κ (green) depositing in the glomerular mesangial region and Coll IV (red) at the GBM. J-L, consecutive sections from the same patient, ×400.

In recent years, some researchers have attempted to apply proteases to treat paraffin-embedded sections to denature cell membranes and retrieve antigenic determinants.35 For example, Fogazzi et al3 examined pathologic samples from patients with IgAN (n = 10), MN (n = 8), and LN (n = 10) and compared the fluorescence intensity in frozen sections and paraffin-embedded sections treated with protease. Their results indicated that the immunofluorescence intensity of several major antigens (IgG in MN, IgA in IgAN, and IgG and C1q in LN) was similar for the 2 methods. However, the method of enzyme digestion is difficult to control in practice. The paraffin-embedded sections are very thin, so it is easy for overdigestion to occur, which may damage tissue structure and the antigens. Nasr et al4 compared IF-F and IP staining on deparaffinized, pronase-treated tissue in 71 renal biopsy specimens representing 12 major renal diseases. Concordant diagnostic findings were obtained in 100% of APGN and LN specimens but only in 88% of IgAN, 60% of MPGN, 50% of MN, and 20% of anti-GBM specimens. Qualman and Keren5 used trypsin to digest paraffin-embedded sections and analyzed 52 samples from patients with various renal diseases. They reported a concordance rate of 80% to 90% with IF-F in the detection of IgG, IgM, and IgA. However, there is still uncertainty with IF-P, and it is not widely accepted in clinical diagnosis.

Shi et al9 reported that a brief microwave heating could increase the immunohistochemical staining intensity of paraffin-embedded biopsy sections and speculated that this was due to exposure of intracellular antigens and recovery of cell activity. Other immunohistochemistry studies have reported that paraffin-embedded sections following microwave retrieval can preserve tissue structure and protect the antigens.1016 Use of conventional antigen retrieval reagents allows visualization of an antibody with a weak signal, thereby increasing the sensitivity of staining.1719 Immunohistochemistry of paraffin-embedded sections following microwave retrieval was reported with satisfactory findings, improving the specific expression of antigen and decreasing the nonspecific expression of antigen.2022 However, few studies have reported applying the microwave antigen retrieval technique for immu-nofluorescence staining in paraffin-embedded sections.

In this study, we found that some plasma proteins may have caused false-positive results after the initial microwave procedure. However, there were no detectable false-positive results after fixation of sections in 4% paraformaldehyde and a second microwave antigen retrieval procedure. We attribute this to the fixation and preservation of antigens by the second retrieval procedure and to an increase of cell activity. In addition, the microwave power was increased from 320 W in the first procedure to 850 W in the second procedure, and all sections were preheated when the water temperature reached 80°C. Therefore, the second microwave antigen retrieval may have alleviated any uneven heating so that all antigens were completely exposed. The microwave treatment can retrieve and fix the antigens, which increases the possibility of detection, so the positive staining of antigens in IF-DMP samples was higher than that of the sections using enzyme digestion. In our study, the concordant observations were 98% for all immunofluorescence staining, complements, κ, and λ; 100% for IgG, IgA, and IgM; 97% for C3; and 96% for C1q between the IF-F and IF-DMP techniques. The concordance of IF-DMP and IF-F in the detection of major antigens in the samples from patients with MN, LN, and IgAN was as high as 100%. In APGN, MPGN, LN, and anti-GBM disease, IF-DMP and IF-F performed similarly in the detection of IgG, C3, and C1q. This indicates that IF-DMP is not inferior to IF-F in the detection of C3 and C1q.

One study reported that the accuracy of diagnosis for glomerular lesions is 65% if the tissue section contains only 5 glomeruli, but it is 95% when a tissue section contains 15 glomeruli.23 Therefore, tissue sections should have at least 15 glomeruli for a renal tissue pathologic examination. We found that the IF-DMP tissue section provided 27.51 ± 6.98 glomeruli, whereas the IF-F section provided only 6.92 ± 1.15 glomeruli. These findings indicate that IF-DMP can give us much more comprehensive and detailed information, which favors the accuracy of diagnosis. Moreover, IF-DMP provided more comprehensive data on glomeruli than did IF-F, a definite benefit for pathologic diagnosis.

The microwave heating of biopsy sections has 2 main effects: fixation and antigen retrieval. So, unlike enzyme digestion, microwave irradiation causes almost no damage to the tissue structure. In addition, double staining of IF-DMP has advantages of higher sensitivity and absence of background staining. IF-DMP provides clear images of tissue structure and location of antigens as precisely as IP staining. IF-DMP can be examined in the same tissue section both by a light microscope and a fluorescence microscope. Moreover, double immunofluorescence staining with IF-DMP allows localization of coexpressed antigens, which is very useful for observing correlations between tissue structure and deposition of immune complexes. It also provides more helpful information for clinical pathologic diagnosis.

In summary, our findings have confirmed that IF-DMP can be used as a routine procedure for immunopathologic examination of human renal tissues following renal biopsy and as an alternative to IF-F pathologic diagnosis.

Footnotes

  • * These authors contributed equally to this work.

  • Acknowledgment: We thank Nancy B. Martin for editing and proofreading the manuscript.

  • This study was funded by grants No. 30630033, No. 30772296, and No.81170312 from the National Natural Science Foundation of China; grant No. 2011CBA01003 from the National Basic Research Program of China (973 Program); and grant No. 2011CB964904 from the National Key Scientific Program of China.

References

View Abstract