Human Pathology (2014) 45, 1766–1772
www.elsevier.com/locate/humpath
Original contribution
An evaluation of Congo red fluorescence for the diagnosis of amyloidosis☆,☆☆ Cecilia G. Clement MD a , Luan D. Truong MD a,b,⁎ a
Department of Pathology and Genomic Medicine, The Methodist Hospital, Houston, TX 77030 Department of Pathology and Laboratory Medicine, Weill Cornell Medical College, New York, NY 10065
b
Received 28 December 2013; revised 23 April 2014; accepted 25 April 2014
Keywords: Amyloid; Congo red stain; Fluorescence; Texas red filter; Birefringence; Polarized light
Summary Congo red stain apple-green birefringence under polarized light is the most popular method for detecting amyloid; however, it has limitations. The goal of this study was to evaluate if examination of Congo red stain by fluorescent microscopy (FM) significantly enhances the diagnostic yield. Congo red–stained tissue sections were retrospectively and prospectively examined by light microscopy (LM) with and without polarizer and by FM using the Texas red filter and results by each method compared. Congo red–stained amyloid recognized by LM was unequivocally and easily identified by FM in each of 48 cases. In 22 of them, FM either confirmed the presence of a small amount of amyloid or lead to a definitive diagnosis, which was otherwise missed. Eight cases with Congo red–negative by LM were also negative by FM. In 8 cases with a false-positive Congo red stain, FM still detected the signal in 5, but it was absent in 3 cases. In conclusion, Congo red fluorescence improves the diagnostic yield of LM for both positive and negative cases. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Amyloidosis comprises a spectrum of diseases characterized by the deposition of amyloid in a variety of organs. These deposits can be local or generalized, the latter causing a progressive disease with poor prognoses. The diagnosis of amyloidosis relies on identification of amyloid deposits on tissue sections. Traditionally, material positive by Congo red stain, which also shows apple-green birefringence under polarized light, is considered the criterion standard for diagnosing amyloid. Indeed, the ☆ Competing interests: The authors have no conflicts of interest to disclose. ☆☆ Funding/Support: The authors have no funding to disclose. ⁎ Corresponding author. Department of Pathology and Genomic Medicine, The Methodist Hospital, 6565 Fannin St, Houston, Texas 77030. E-mail address:
[email protected] (L. D. Truong).
http://dx.doi.org/10.1016/j.humpath.2014.04.016 0046-8177/© 2014 Elsevier Inc. All rights reserved.
presence of apple-green birefringence is even mandatory for this diagnosis in some studies [1-4]. However, this method has its limitations. It has been shown to be unreliable in some cases, in particular, with a minute amount of amyloid [2,5,6]. Moreover, the Congo red dye bound to amyloid needs tissue sections to be of a required thickness (N8 μm) to show birefringence [7]. Therefore, if birefringence is solely relied on for the diagnosis of amyloidosis, some cases would be missed. More than 4 decades ago, Klatskin [8] questioned the specificity of the Congo red polarization method; under appropriate processing conditions, he found foci of green birefringence in connective tissues of liver and several normal organs both in man and in the rat. And even the “apple-green birefringence in polarized light,” words traditionally use to describe the properties of Congo red– stained amyloid, has been questioned more recently by Howie et al [9], who observed that more than one color may be seen, and in some cases, green is not even present.
Congo red fluorescence in amyloidosis To increase the detection of amyloid, alternative methods have been used, including immunohistochemical staining using antibodies against variable amyloid components or examination of Congo red–stained tissue sections by fluorescent microscopy (FM). Although combined Congo red stain and immunohistochemistry has been reported to be more sensitive than Congo red stain alone [10], limitations of the immunohistochemical method may include availability of antibodies and the possibility of negative results in cases with very small amounts of amyloid [6]. There is growing evidence supporting the high sensitivity of Congo red fluorescence to detect amyloid [1,5,6,11,12]. The goal of this study was to evaluate the possibility that examination of Congo red–stained tissue sections by FM significantly enhances the diagnostic yield in histologic sections from a variety of tissue types.
2. Materials and methods Sixty-four specimens were studied in both retrospective and prospective manner. This study was reviewed and approved by the Institutional Review Board of The Houston Methodist Hospital, Research Institute, Houston, TX. These cases were categorized into 3 groups. Group 1 included 26 cases with abundant amyloid deposition, in which the identification of amyloid was straightforward, even by routine hematoxylin and eosin (H&E) stain; this group included 6 heart specimens (3 explants and 3 biopsies); 5 liver specimens (4 biopsies and 1 explant); 3 kidney biopsies; 3 corneal excisions; 2 small bowel biopsies; 2 pancreatectomy specimens; and 5 excisional biopsies of thyroid, urinary bladder, lymph node, vocal cord, and lung (one of each). Group 2 consisted of 22 cases in which amyloid deposition was scant or minuscule; in some of these cases, the detection of amyloid was initially problematic but finally confirmed. These cases included 9 gastrointestinal tract (esophagus, stomach, duodenum, and colorectal) biopsies; 4 liver specimens (3 biopsies and 1 excision); 3 heart biopsies; 3 fine needle biopsies of abdominal fat pad; and biopsies (one of each) of kidney, tongue, and liver. In each case of this group, tissue was submitted with the clinical suspicion for amyloidosis. Tissue was stained with Congo red stain for each of them. In several cases, amyloid was present in small amount but was still easily recognizable by light microscopy (LM). In other cases, the amyloid was so scanty that it was initially missed by LM (examined by the attending pathologists). These cases were then consulted with the authors, who examined by the same tissue sections by FM. This demonstrated illumination typical for amyloidosis. Examination by LM again focusing on the areas of illumination now convincingly showed amyloid, albeit in very scant amount. The diagnosis of amyloid in each of these cases was also supported by the presence of amyloid in other
1767 organs and/or laboratory finding including the presence of circulating monoclonal light chains. Group 3 included 16 cases with tissue types that may simulate amyloidosis in routine H&E stain, some of which are also well known for a false-positive Congo red staining. They included elastic tissue, poorly cellular hyalinized fibrous tissue, colloid of thyroid gland, aggregated red blood cells in vascular lumen, uromodulin deposition in kidney tissue, or fibrin deposition. The correct nature of these tissue types in most cases were readily recognized by LM morphology. In others, additional stains such as Verhoff’s stain for elastic tissue, Masson trichrome stain for fibrous tissue and fibrin, or periodic acid-Schiff stain for uromodulin helped identify the tissue types. In each case, tissue sections of 4- to 10-μm thick were previously or prospectively stained for Congo red. Congo red staining was performed using the alkaline Congo red method for amyloid, as described by Puchtler et al [13]. H&E-stained and Congo red–stained sections of each case were examined by routine LM, with or without a polarizer and by FM (Olympus microscope; Olympus America, Center Valley, PA) under a fluorescein isothiocyanate filter (FITC) with an excitation peak at 490 nm and a fluorescence emission maximum of 520 nm and a Texas red filter with an excitation peak of 596 nm and emission maximum of 620 nm. In some selected cases, unstained slides consecutive to those used for Congo red stain were also examined in the same way. The results by each method were compared.
3. Results The Congo red–stained amyloid deposits showed the typical pinkish to pink-reddish color in bright light. When Texas red filter was used with ultraviolet (UV) light, the Congo red–stained amyloid showed a bright red color against a dark background. The unstained, H&E-stained, and the Congo red–stained slides, regardless of the presence or absence of amyloid, showed no specific staining pattern under FITC filter. Amyloid was strongly Congo red–positive by LM examination in all 26 cases of group 1. The “apple-green” birefringence expected by examination with polarizer was very focal or even absent in some cases (Fig. 1A and B), or it was focally identified in areas without congophilia, as seen in Fig. 1D and E. Examination using the fluorescent Texas red filter showed strong signal in all Congo red–positive amyloid cases (Fig. 1C, F, H, and J). Structures other than amyloid deposition did not show this intense type of illumination (Fig. 1D-F). For the 22 cases of group 2, similar observations as in group 1 were made. In each case, the scant amount of Congo red–stained amyloid was easily and instantaneously recognized under fluorescent Texas red filter examination. Furthermore, there were 2 cases (myocardial needle biopsy
1768
C. G. Clement, L. D. Truong
Fig. 1 Comparison of Congo red–stained amyloid deposits under LM, FM, and polarized light (PL). Amyloid deposits in small bowel submucosal vessel and surrounding Brunner glands (A-C) show strong congophilia by LM (A), which is much enhanced by FM (C); in contrast, PL shows very focal birefringence (B). D to F, Amyloid deposits in fibroadipose tissue, which was included in a liver biopsy. FM provides a more intense amyloid signal compared with LM versus reducing the background signal (D and F); under PL, only small areas of amyloid are illuminated, whereas the birefringence is mainly seen in nonamyloid elements (E). G and H, Abundant glomerular amyloid deposits clearly shown by both LM and IF. I and J, Abundant liver amyloid deposit encircling hepatocytes clearly shown by both LM and IF (original magnification, ×200 for all panels).
and fine needle biopsy of abdominal fat pad) in which absence of amyloidosis was formally diagnosed, but retrospective examination of the Congo red–stained sections by Texas red filter clearly identify amyloid
deposition, which previously escaped careful examination by routine LM (Fig. 2). Within group 3, Congo red stain was negative in 8 cases in which H&E-stained sections were suspicious for amyloid,
Congo red fluorescence in amyloidosis
1769
Fig. 2 Comparison of Congo red–stained small amyloid deposits under LM (A, C, E, and G) and FM (B, D, F, and H). FM reveals amyloid deposits in a small vessel of a heart biopsy (A and B), surrounding myocardial fibers (C and D), in a fine needle aspiration of abdominal fat pad (E and F), and in a small vessel of rectal submucosa (G and H). In each of these cases, the amyloid deposit was initially missed or in doubt by LM (original magnification, ×400 [A, B, E, and F], ×200 [C, D, G, and H]).
and FM with Texas red filter was also negative in all of them. The remaining 8 cases in this group showed focal congophilia of variable but often weak intensity in nonamyloid elements. These elements included red blood cells, thyroid colloid, and poorly cellular fibrous tissue. FM eliminated the nonamyloid Congo red–positive signal in 3 of
these cases, thus improving the differentiation of amyloidspecific from nonspecific signal (Figs. 1D-F; 3A and B). However, FM showed positivity in the other 5 cases but with such a subdued degree in most cases, compared with typical amyloidosis, that differentiation from amyloid signal was much facilitated (Fig. 3C-F).
1770
C. G. Clement, L. D. Truong
Fig. 3 Visualization of nonspecific Congo red signal of colloid in thyroid tissue (A and B) and red blood cells in glomerular capillaries (C-G). Nonspecific signal in thyroid colloid under LM (A) is virtually eliminated by FM, whereas Congo red–stained amyloid deposit is much enhanced (B). The false-positive Congo red stain of red blood cells in glomerular capillaries by LM (C) is still detected by FM (D). The false-positive Congo red stain of red blood cells in glomerular capillaries in another case (E) is attenuated by FM (F); red blood cells shown by H&E stain in this glomerulus (G) (original magnification, ×200 [A and B]; ×400 [C-G]).
4. Discussion Since Benhold first reported the value of Congo red stain in detection of amyloid in early 1920s, this stain has undergone several modifications to increase its sensitivity
and specificity [1,7]. The alkaline Congo red method of Puchtler et al [13] is currently the most widely used in laboratories of the United States. Congo red–stained amyloid fluorescence was described as early as 1959 by Cohen et al [14], who emphasized that
Congo red fluorescence in amyloidosis “when carried out meticulously, it was useful in delineating small accumulations of amyloid.” Congo red dye shows a fluorescent activity when bound to amyloid fibrils, and this is visualized under UV light. As a fluorochrome, Congo red dye reemits light upon light excitation. The excitation and/or emission filters should match the spectrum of excitation and emission wavelengths of the fluorochrome used to stain specimens; otherwise, the fluorochrome staining product would be affected. Congo red dye bound to amyloid is easily seen as red fluorescence using a filter for an excitation wavelength in the range 530 to 585 nm and emission at more than 600 nm [15], suitable for fluorochromes such as Texas red. Therefore, we obtained excellent results using a Texas red filter, whereas there was no specific staining pattern under FITC filter, with excitation/emission maxima below the Congo red range. Descriptions of the increased sensitivity of Congo red FM for detection of the smallest amounts of amyloid have been alluded since early 1960s, remarking the much more intense staining effect of FM compared with bright light [2,11]. More recent studies also supported the concept of Congo red fluorescence as the most sensitive method for detection of amyloid. Linke [6] compared FM with Congo red staining method alone and combined with immunocytochemistry. He found that Congo red staining was the least sensitive method for detection amyloid, followed by immunocytochemistry, and Congo red FM being the most sensitive method, even in cases of small deposits. Giorgadze et al [1] reported that using FM in Congo red–stained fat pad fine needle aspiration smears, there was an improvement in the percentage of positive results from 15.4% under polarized light to 39.2% by FM examination. In a retrospective study of Congo red– stained renal biopsies by Sen and Basdemir [5], after reevaluation by FM, 87 cases initially positive under bright light were confirmed by FM, and one new case was identified. The sensitivity and specificity of Congo red stain for amyloidosis have been subjected to several studies [1,46,8,10]. The current study showed that, in each of the 26 cases with abundant Congo red–positive amyloid deposition, FM readily confirmed the finding. Furthermore, in 22 cases with scant amyloid, FM also not only readily confirmed the finding, but in at least 2 of these cases also revealed amyloid, which was originally missed by LM examination. Retrospective examination of the same tissue sections, after amyloid was recognized by IF, indeed revealed scant but convincing amyloid deposition. These observations clearly indicate that FM facilitates the recognition of amyloid (better sensitivity) (Fig. 2). The specificity of Congo red stain is a controversial topic. It has been well documented that even the best version of Congo red stain may yield a false-positive reaction, reflecting a lack of specificity [8]. Many authors still consider that the specificity of the Congo red staining procedure for diagnosis of amyloid depends on demonstration of apple-green birefringence. However, tissue elements such as collagen, elastin, and others, may display birefringence of varying colors that may be
1771 misinterpreted as amyloid. Therefore, one goal of the current study is determining whether FM would alleviate this problem. In this aspect, we have selected 16 cases with false-positive Congo red stain and compared the results of FM examination and polarized light examination, a traditional method to differentiate amyloid (birefringence) from its simulators (no birefringence). Examination under polarized light showed that apple-green birefringence was noted in 60% of the cases in group 1 and 35% of cases in group 2. Birefringence (albeit of variable color) was also seen in 50% of cases in group 3. These observations suggest that, at least in our hand, birefringence of Congo red, stain may not be sensitive or specific for confirming the diagnosis of amyloidosis, an observation previously reported. It has been emphasized that Congo red birefringence may depend on several factors including the thickness of the tissue section, the intensity of the light source, and the amount of amyloid deposition. These factors may explain some of the aberrant observations in the current study. Comparing polarized light examination and FM in cases with false- positive Congo red stain showed that in 8 of 16 of these cases, birefringence is still detected, whereas FM showed illumination in only 5 of 16. These observations indicate that FM examination does not eliminate the problem of falsepositive Congo red stain, but it does improve the specificity of this stain. This is an important matter in the diagnostic utility of FM that may need further study. A caution is therefore essential to avoid incorrect interpretation of the FM findings in the context of false-positive Congo red stain. Also pertinent is the observation that if there is no Congo red stain, FM examination will not reveal any illumination. Congo red–positive amyloid deposition in large amount is readily and accurately recognized by LM. However, in problematic or difficulty cases, especially those with scant amyloid in small biopsy, FM examination is an important adjunct to the diagnosis. On the other hand, as mentioned previously, the problem of false positivity, although improved, is not eliminated by this technique. The current data do not allow us to develop objective conclusions on the comparative utility of thioflavin T and Congo red fluorescence for the diagnosis of amyloidosis. However, an earlier study suggested that both are sensitive methods and able to detect even small amounts of amyloid [2]. Laser microdissection combined with mass spectrometer has recently emerged a novel technique for detection and, more importantly, classification of amyloidosis. This technique involves LM identification of foci of tissue amyloid deposition by Congo red stain. These foci are then dissected out by laser microscope, enzymatically digested into peptide components, which are subjected to mass spectrometric analysis. The obtained peptide profiles will enable accurate identification of the type of amyloidosis. Because there are a large number of chemical types of amyloidosis (approximately 25 currently) and immunohistochemical staining for these types of amyloid protein in tissue sections is often problematic, this novel technique is essential for amyloid classification. And it achieves this goal very well. Yet, its success remains
1772 dependent on correct identification of amyloid deposit in tissue and the presence of adequate amount of amyloid for analysis. This technique therefore does not obviate the need for accurate tissue diagnosis of amyloidosis. In summary, the current study suggests that Congo red fluorescence has several diagnostic advantages. It improves both sensitivity and specificity of the traditional Congo red stain under regular or polarized light. This technique also has the advantage of its simple use because it uses the same Congo red–stained slide, and only a UV light illumination is required. We encourage an integrated approach, using the traditional Congo red stain along with FM, in all cases with a suspicion of amyloidosis because it will significantly increase the diagnostic yield of Congo red stain for both positive and negative cases.
References [1] Giorgadze TA, Shiina N, Baloch ZW, Tomaszewski JE, Gupta PK. Improved detection of amyloid in fat pad aspiration: an evaluation of Congo red stain by fluorescent microscopy. Diagn Cytopathol 2004;31:300-6. [2] Cooper JH. An evaluation of current methods for the diagnostic histochemistry of amyloid. J Clin Pathol 1969;22:410-3. [3] Getz MA. The classification and typing of amyloid deposits. Am J Clin Pathol 2004;121:787-9.
C. G. Clement, L. D. Truong [4] Marcus A, Sadimin E, Richardson M, Goodell L, Fyfe B. Fluorescence microscopy is superior to polarized microscopy for detecting amyloid deposits in Congo red-stained trephine bone marrow biopsy specimens. Am J Clin Pathol 2012;138:590-3. [5] Sen S, Basdemir G. Diagnosis of renal amyloidosis using Congo red fluorescence. Pathol Int 2003;53:534-8. [6] Linke RP. Highly sensitive diagnosis of amyloid and various amyloid syndromes using Congo red fluorescence. Virchows Arch 2000;436: 439-48. [7] Elghetany MT, Saleem A, Barr K. The Congo red stain revisited. Ann Clin Lab Sci 1989;19:190-5. [8] Klatskin G. Nonspecific green birefringence in Congo red-stained tissues. Am J Pathol 1969;56:1-13. [9] Howie AJ, Brewer DB, Howell D, Jones AP. Physical basis of colors seen in Congo red-stained amyloid in polarized light. Lab Invest 2008;88:232-42. [10] Linke RP, Gartner HV, Michels H. High sensitive diagnosis of AAamyloidosis using Congo red and immunohistochemistry detects missed amyloid deposits. J Histochem Cytochem 1995;9:863-9. [11] Puchtler H, Sweat F. Congo red as a stain for fluorescence microscopy of amyloid. J Histochem Cytochem 1965;13:693-4. [12] Rocken C, Sletten K. Amyloid in surgical pathology. Virchows Arch 2003;443:3-16. [13] Puchtler H, Sweat F, Levine M. On the binding of Congo red by amyloid. J Histochem Cytochem 1962;10:355-64. [14] Cohen AS, Calkins E, Levene CI. Studies on experimental amyloidosis. Analysis of histology and staining reactions of caseininduced amyloidosis in the rabbit. Am J Pathol 1959;35:971-89. [15] Fowler DM, Koulov AV, Alory-Jost C, Marks MS, Balch WE, Kelly JW. Functional amyloid formation within mammalian tissue. PLoS Biol 2006;4:100-7.
