DIRECT IMMUNOFLUORESCENCE AND IMMUNOHISTOCHEMISTRY IN DIAGNOSTICS OF GLOMERULONEPHRITIS

Th e needle biopsies from  transplanted and native kidneys have been processed and a prospective analysis of pattern, intensity and distribution of immunoglobulin deposits (IgA, IgG and IgM) and complement components (Cc and Cq) identifi ed in these lesions has been carried out by immunohistochemistry with three step immunoperoxidase, in the period from  to . Th ose deposits were previously detected and analyzed by immunofl uorescence. Th e samples consisted of  renal biopsies, previously diagnosed with glomerulonephritis and positive immunofl uorescence and  renal biopsies without morphologic changes and deposits on immunofl uorescence. , of the analyzed samples showed the identical results of the deposits of immunoglobulin and components of the complement with both, immunohistochemistry and immunofl uorescence method. Sensitivity of the immunohistochemistry method with three step immunoperoxidase for all analyzed immunoglobulin and complement components is high (,), while specifi city for the same method is ,. Standardized method of the three step immunoperoxidase on the paraffi n embedded, formalin fi xed needle renal biopsies could successfully replace the immunofl uorescence method in diagnostic of GN, with the emphasis on a follow up and control of each single step in the procedure of the method.


Introduction
The approaches to morphologic analysis of renal biopsies are limited, as is the number of pathognomic morphological changes.One of the key elements for diagnosis of glomerulonephritis is the finding of immunoglobulin deposits and deposits of complement components exclusively analyzed by the method of direct immunofluorescence (DIF), which has been a gold standard for diagnosis of immunological lesions on renal biopsies for half a century ().Even though it is technically relatively simple laboratory method, it is not an easy task to perform with high quality results.Th is method is also highly sensitive and specifi c, although the fluorescence is not permanent, and its documentation is only possible with quality microphotography, which is not always available.For immunofl uorescence analysis, in notable number of cases "locating" fl uorescent immune deposits in the context of morphological picture is complicated, due to hardly noticeable signal.Th e reasons mentioned above led to development and use of immunohistochemistry method on paraffi n tissue sections, instead of DIF method (,).If immunohistochemistry method would allow satisfactory view of immune deposits, a form of microscopic diagnostics in which immune deposits could be viewed in context of clearly discernable and quality background, would be provided.At the same time, stable slides would be secured, for further analysis of immune deposits ().Th e objective of our work was to analyze distribution, pattern and intensity of immunoglobulin (IgA, IgG, IgM) and complement components (Cc and Cq) on renal biopsies using immunohistochemistry method of three step immunoperoxidase, and compare sensitivity of this method to the method of direct immunofl uorescence.

Renal tissue samples
Th e paper represents retrospective-prospective study which involves detection and analysis of distribution, pattern and intensity of immunoglobulin deposits (IgG, IgA, IgM) and complement components (Cc and Cq) by immunohistochemistry method of a three step immunoperoxidase in the  renal biopsies from the archive of Department of Pathology, Polyclinic for laboratory diagnostics of University Clinical Center in Tuzla, in the period from  to .Previously performed detection of immune deposits done by immunofl uorescence method on the same biopsy samples represents control group.Out of  examined,  biopsies were chosen by the random selection method from the pool of  biopsies, with previously diagnosed glomerulonephritis and confi rmed positive result deposits of immunoglobulin and complement components by immunofl uorescence method.A second group includes  biopsies without apparent glomerular morphological changes and with the absence of immunoglobulin and complement components deposits analyzed by immunofluorescence method.All of the biopsy samples were adequately fi xed and contained enough renal tissue for additional analysis.

IF Technique
All of the tissue samples have been transported to Department of Pathology in physiological solution, according to standard operating procedure for renal biopsies and treated with method of cold fi xation upon arrival.Antibodies labeled with fluorescein-isothiocyanate (FITC) against human immunoglobulin (IgA, IgG, IgM) and complement components (Cc and Cq) were applied on histology sections μm thick (Table).All antibodies are manufactured by DAKO, Glostrup, Denmark.Slides are analyzed by microscope Diastar Reichert-Jung with an appropriate fi lter for FITC fluorescence.A pattern and distribution of immune deposits were analyzed qualitatively, while the intensity was determined semi-quantitatively by  degree scale from  to +, as follows: complete absence of deposit (), mild (+), moderate (+), strong (+) intensity.Upon completion of DIF analysis tissue cylinders were fixated in  buffered formaldehyde (pH ,-,), paraffin embedded, and histology sections μm thick were stained with hematoxylin eosin method for the analysis of basic light microscopic morphological marks and following special staining; Periodic acid-Schiff reagents (PAS), Trichrom-Masson, Van Gieson-Weigert, Impregnation with silver according to Jones.Samples are analyzed with Diastar Reichert-Jung light microscope with the goal of determining the type and stage of morphologic changes of all of the components of renal tissue.Classifi cation was done according to the criteria of WHO for glomerular diseases and Banff  work classifi cation for renal transplantation pathology (,).
* Total protein concentration .Histology slides were mounted on appropriate carriers and placed in center for immune-staining Sequenza-Shandon where all stages of incubation were performed.Pre-incubation stage of  minutes with  normal fetal calf serum is followed with procedure of three step immunoperoxidase.In the fi rst stage slides are incubated with primary antibodies in duration of  minutes.Following antibodies were used: IgA, IgG, IgM, Cc and Cq, all manufactured by DAKO, Glostrup, Denmark (Table ).In the second stage incubation in duration of  minutes is performed with biotin labeled anti-rabbit antibody.Incubation in the third stage is done with streptavidine labeled peroxidase in duration of  minutes.Washes between incubation are done with phosphate buffered saline.
Peroxidase activity is developed with , diaminobenzidine tetrachloride solution and HO substrate.Contrast staining is done with hematoxylin, and upon dehydratation slides were mounted with Canadian balsam.Evaluation of technical quality of slides treated with immunohistochemistry method of three step immunoperoxidase is performed by light microscope.The quality of specimen is evaluated as unsatisfactory in cases with present non specifi c positive staining of adhering serum proteins in glomerular and peritubular capillaries, consequence of poor proteolysis and indirectly an indication of inadequate antigen retrieval (,).Samples observed under light microscope exhibiting tissue damage in pattern of dissolution of cytoplasm and/or cell nuclei as a consequence of an excessive digestion are also evaluated as samples of inadequate quality.In both cases analysis has been repeated on the new sections of the same tissue samples.
Immunoglobulin and complement components deposits were analyzed under light microscope Olympus BX, qualitatively and semi-quantitatively, according to the same module that is used in the method of direct immunofl uorescence (distribution, pattern and intensity), and results are compared to written description of fi ndings performed by immunofl uorescence method.

Statistical analysis
All analyzed samples are included in x table of paired samples in which immunofluorescence method was regarded as the reference test.Statistical analysis has been calculated with the continuity correction included and referred to the chi-square distribution for probability with df (McNamara test).

IgA
In the  (,) biopsy samples of the examined group, fi nding of IgA deposit by IP was identical to DIF fi nding according to intensity, distribution and pattern of deposits (Figure ).In  (,) cases intensity of deposits found by IP was more expressed (Figure ), in  (,) case DIF found deposits were of stronger intensity, while the pattern and distribution of deposits inside glomerules were identical .In  () cases presence of IgA deposit in biopsy samples was observed by IP with a negative immunofl uorescence fi nding, while  (,) case had a positive immunofl uorescence fi nding deposits of this immunoglobulin were not observed by IP.Statistical analysis has shown high specifi city (,) and sensitivity (,) of IP method.Predicted value of positive test for IP method was ,.Statistically signifi cant diff erence was not found in predictions of these two methods (Chi=,, P=,).

IgG
In total of  (,) renal biopsies fi nding of IgG depos- it by immunohistochemistry method of three stage immunoperoxidase was identical to immunofl uorescence fi nding according to intensity, distribution and pattern of deposit.In  () cases IP method showed deposits were more expressed according to intensity, while in  (,) case DIF deposits were of stronger intensity, with identical distribution inside glomerules and pattern of deposit.In  (,) cases IP method showed presence of IgG deposit in biopsy samples with negative DIF fi nding, and in  () cases with positive DIF fi nding, IP deposits of this immunoglobulin were not observed.Statistically signifi cant diff erence was not found in predictions of these two methods (Chi=, P=).Statistical analysis showed high specifi city (,) and sensitivity (,) of immunohistochemistry method.Predictive value of positive test for immunohistochemistry method was ,.

IgM
In total of  (,) renal biopsies fi nding of IgM deposit by immunohistochemistry method of three stage immunoperoxidase was identical to fi nding of immunofl uorescence method according to intensity, distribution and pattern of deposit (Figures ,).Immunohistochemistry evidence of deposits was of stronger intensity in  () cases, while in  (,) samples DIF method detected deposits of stronger intensity.In  () cases IP detected IgM deposits on samples with previous negative DIF fi nding.Statistically signifi cant diff erence in predictions of DIF and method of three step immunoperoxidase was not found (Chi=,, P=,).Statistical analysis indicates high specifi city (,) and sensitivity () of IP method.Predictive value of positive test for immunohistochemistry method is ,.

Cc
In  () analyzed biopsies, finding of Cc deposit and complement components, finding of IP method was identical to DIF finding according to intensity, pattern and distribution of deposits (Figures ,).In  (,) case intensity of deposits detected by IP had slightly stronger expressed.In  (,) cases IP method demonstrated presence of Cc in biopsies which had negative DIF fi nding, and  (,) case with positive DIF fi nding IP method did not detect deposit of this complement component.Statistically significant diff erence was not found in predictions of these two methods (Chi=,, P=,).High sensitivity (,) and specificity (,) of method of three step immunoperoxidase was determined with statistical analysis.Predictive value of positive test for this method is ,.

Cq
In  (,) biopsy samples of both groups, immunohistochemistry fi nding of deposit Cq complement com-ponent was identical to fi nding of DIF (Figure ), and in all cases with present deposits they had identical intensity, pattern and distribution.In  (,) cases analyzed by IP method presence of deposit Cq was found in biopsy in which DIF fi nding was negative.All detected deposits in these biopsies had intensity of + and granular shape localized in the wall of glomerular capillaries (Figure ).In predictions for the two compared methods statistically signifi cant diff erence (Chi=,, P=,) has been found.Th e evidence is high sensitivity () and low specifi city (,) of immunohistochemistry method of three step immunoperoxidase.Predictive value of a positive test for immunohistochemistry method is ,.

In general
A comparison of the results for all analyzed immunoglobulin and complement components established that in  (,) samples fi ndings were identical for both methods.In  (,) renal biopsies with positive fi nding with both methods, deposits of various intensities are detected.In  (,) tissue sample with previously negative DIF fi nding IP method detected deposits of all immunoglobulin and complement components.In  () cases IP did not confi rm presence of deposits in biopsy samples with previously positive immunofl uorescence.Results acquired by comparison of the two methods for all analyzed antibodies are shown in Tables  and .
Statistically signifi cant diff erence was found in prediction of immunofluorescence method and method of three step immunoperoxidase (Chi=,, P<,).Statistical analysis showed high sensitivity (,) of immunohistochemistry method, while specifi city is ,.
Predictive value of a positive test for this method is ,.

Discussion
According to our knowledge and Molne report, only four studies, which compared immunofl uorescence and immunohistochemistry method results of immune deposits detection in renal biopsies in a way to allow statistical comparison with our results, have been reported.() In three studies from earlier periods the authors compared results of DIF-e with immunohistochemistry method peroxidase-antiperoxidase, while Molne in his work used method of immunoperoxidase with Dako En Vision HRP system (,,,).Considering that basic postulate for acquiring quality slides by immunohistochemistry method are standardized procedure and educated and experienced laboratory personnel, it was decided for the method of three step immunoperoxidase with streptavidine, which has been used in our lab in diff erent tissue samples over one decade.Th is will possibly lead to eventual introduction of this method as a routine diagnostic procedure of renal disease in the future.Identical results acquired with comparison of DIF and immunohistochemistry method which in studies of MacIver, Sinclair and Howie is from  do  is higher from the value we obtained (,), while the value from the study done by Molne and associates is somewhat lower (,) (,,,).Discrepancy between these results done by immunofl uorescence and immunohistochemistry method in our and studies mentioned above is a consequence of distinction of application of criteria for determining uniformity of positive fi ndings.We considered identical only fi ndings in which deposits were of same intensity, while the same criteria was not taken in to consideration in compared studies.Statistical analysis in total results of the studies done by McIver and Sinclair did not found signifi cant diff erence between DIF and IP in detection of immunoglobulin and complement components (,).In studies done by Howie, Molne and in our study signifi cant diff erence between these two methods has been observed (p<,) (,).
Looking independently at the results of individual immunoglobulin and complement components in both mentioned studies for IgA and IgG signifi cant diff erence was not found, although our study found noticeable and statistically signifi cant diff erence only in results of Cq complement components.This significant difference, looked at in whole is a consequence of found diff erences in analysis of individual antibodies in the two compared methods.In the study done by Howie and associates Cc is more often detected with fl uorescent method, while in Molne's study Cq and IgM are more often detected by immunohistochemistry method (,).In our study, all analyzed immunoglobulin and complement components are detected in certain number of renal biopsies with previously negative immunofl uorescence fi nding.
In the study done by Turner and associates in  comparison of deposit fi ndings for IgA, IgG, IgM and Cc by immunofl uorescence and immunoperoxidase method in  of renal biopsies expressed uniformity in  to  cases ().Diff erence in relation to percentage of identical fi ndings in our study is caused by the fact that in mentioned study Cq complement component was not analyzed, while our study reported decrease in percentage of identical findings (,).
In the study done by Jackson and associates in , in  renal biopsies immunofl uorescence method is compared with immunohistochemistry method of immunoalkaline phosphatase, with the diff erence that in our study alkaline phosphatase is used as a substrate for deposit labeling ().Uniformity of  for all compared antibodies (IgA, IgG, IgM, Cc, Cq and fi brinogen) is diff erent from our results, probably a consequence of the diff erence in compared antibodies, due to the fact that in our study we did not analyze presence of fi brinogen deposits.
In the reference to the role of complement activation in etiopathogenesis of glomerulonephritis, deposits of complement components have a diagnostic role, especially in determining glomerular damage in the context of systematic diseases.In our study in , cases we detected Cc deposits in biopsies with previously negative  ), it is more probable that during tissue preparation for immunofl uorescence some factors are relocated or removed, which disables their retrieval and detection with this method.On the other hand, IP technique allows detection of antigens lost in the process of tissue preparation for immunofluorescence analysis, especially in the process of tissue washes.
According to the research done on immunohistochemistry staining of renal tissue, especially for observation of immunoglobulin and complement components in deposits of immune complexes, the most eff ective pretreatment is the one with proteolytic digestion ().Duration of digestion must be adjusted to the stage of fi xation and thickness of histology sections which need to be constant.In our study sections paraffin embedded tissues previously fi xed in  formaldehyde buff er were used.Duration of fi xation process also needs to be standardized, depending on specimen dimensions, because it enables fast and uniform process of fixation, as a precondition for controlled antigen retrieval.

Conclusion
Th ree step method of immunoperoxidase is sensitive and specifi c in determining deposits of immunoglobulin and complement components in renal biopsies, and it has certain advantages when compared to immunofl uorescence method ().Considering diff erent results, in the use of this method, authors recommend parallel use of immunofl uorescence and immunoperoxidase method as a way of results verifi cation.According to the results of our study in the cases when the archive material was accessible and previous process of fi xation standardized, retrospectively prospective study can exclude need for long lasting parallel use of two methods.Standardized method of three step immunoperoxidase can represent progress toward elimination of immunofl uorescence as a most signifi cant method in detection of immune complex in renal biopsies.Each procedural step, including starting preparation of tissue sample, has to be controlled and standardized, for the purpose of getting reliable results.Excellent technical advancement in immunohistochemistry methods, wide use in all areas of pathology and well trained laboratory personnel will enable use of three step immunoperoxidase as an equivalent method in analysis of immune deposits in renal biopsies, especially in case when fresh tissue is not available for immunofl uorescence.

TABLE 1 .
Antibodies used for DIF pretreated in citrate buff er( mmol/dm, pH=,)at °C in duration of  minutes.Next step is treatment of histology slides with avidine-biotin blocking solution in duration of  minutes, two times at room temperature (Dako Cytomation, Carpinteria, USA)

TABLE 3 .
Presence and intensity of immunoglobulin and complement components deposits

TABLE 4 .
Results of immunofl uorescence and immunohistochemistry method :