A Novel Histochemical Method for a Simultaneous Staining of Melanin and Collagen Fibers

Abstract

For the histopathologic diagnosis of melanocytic lesions, it could be necessary to identify the melanin pigment because its visualization is unspecific with hematoxylin-eosin (HE). The Fontana-Masson (FM) technique is used in histopathology in this type of lesion, which allows the identification of the pigment, but it loses all the morphologic parameters. The authors describe a modification to the FM method, for the evaluation of the morphology, the argentaffin reaction of the melanin, and collagens fibers of the extracellular matrix simultaneously, for which they have developed the Fontana-Masson picrosirius (FMPS) method. Biopsies of different melanocytic lesions were used for the performance of the HE, FM, and FMPS methods. The pixel intensity of the reaction for melanin, collagen, and epithelium was determined with ImageJ software. The FMPS method allows the evaluation of morphological characteristics, identifying the melanin pigment and collagen fibers with high intensity simultaneously. This method does not differ significantly from FM in the identification of melanin, maintaining its sensitivity and specificity. In addition, it does not differ in the demonstration of the morphology with HE. However, FMPS is significantly superior in the identification of collagen fibers. The FMPS method combines morphological and histochemical parameters that could be useful in the study of pigmented lesions of melanocytic origin.

Melanin is produced by melanocytes of the melanoepidermic unit and other cells types. These cells secrete and distribute the melanin pigment, which provides protection from ultraviolet radiation (Passeron et al. 2007; Hearing and Leong 2006). This pigment consists of anionic molecules with high molecular weight, susceptible to degradation by strong bases. It is formed by the oxidation of tyrosine in 3,4-dihydroxyphenylalanine (DOPA), which is converted to melanin. These enzymatic reactions are performed in units surrounded by membranes called melanosomes, which are transferred to keratinocytes to perform their function (Ross and Pawlina 2005; Hearing and Leong 2006). This pigment is argentaffin, which has the capacity to reduce metallic solutions directly (Bancroft and Gamble 2008).

Several lesions of melanocytic origin can be classified as benign or malignant (Stevens and Lowe 2000; Houghton and Polsky 2002). A heterogeneous group of melanocytic proliferative congenital and acquired lesions has clinical, morphological, histochemical, and immunohistochemical features in common (Murali et al. 2009). Melanoma is one of the most important malignancies characterized by high invasiveness and metastasis (Srivastava et al. 2003; Hofmann et al. 2000). The differential diagnosis of these lesions is usually morphological and immunohistochemical. However, the identification of melanin in the pigment is necessary for a correct diagnosis (Murali et al. 2009; Stevens and Lowe 2000).

Visualization of melanin is nonspecific with hematoxylin-eosin (HE) staining and becomes difficult when the pigment is scarce or other pigments are present. For this reason, histochemical techniques are used for melanin identification, such as the DOPA oxidase, ferrous iron, and Fontana-Masson (FM) methods (Bancroft and Gamble 2008; Kiernan 2008). The FM method is most often used in histopathology for the study of this lesion (Bancroft and Gamble 2008). This method is based on the reduction of ammoniacal silver to metallic silver due to the melanin in the melanosomes. The product of the reaction is an insoluble black precipitate, which enables light microscopy identification (Bancroft and Gamble 2008).

Although the FM method is effective for the demonstration and identification of the melanin pigment (Bancroft and Gamble 2008; Kuznitzky et al. 2003), the evaluation of morphological parameters is very difficult when this method is applied, and in most cases, it is not possible to make a comprehensive assessment of the melanocytic lesion. There are cases of difficult diagnosis because of the undifferentiation of neoplastic cells, poor pigmentation, and sometimes the presence of other pigments. In the study of this type of lesion, the characteristics of collagen bands, neoplastic cell organization, and the pigmentation degree should be evaluated for establishing the diagnosis (Murali et al. 2009; Ferrara et al. 2005). Different techniques should be used and interpreted in parallel, and a correlation of these findings must be further performed. Therefore, for the evaluation and diagnosis of pigmented melanocytic lesions, the development of techniques that allow a comprehensive assessment of lesions in a unique histological slide would be useful.

In this work, we describe the new Fontana-Masson picrosirius (FMPS) histochemical method based on conventional FM and picrosirius histochemical methods, for the simultaneous evaluation of morphological characteristics, the argentaffin reaction of the melanin pigment, and histochemical identification of stromal collagen fibers.

Materials and Methods

The FMPS method was applied on 52 biopsies with several lesions of melanocytic origin (including 18 cases of reed nevi, 15 cases of blue nevi, 5 cases of melanoma, 5 cases of melanocytic intradermal nevus, 2 cases of dysplastic nevus, 2 cases of Spitz nevus, 2 cases of junction nevus, 2 cases of normal skin, and 1 case of compound nevus) obtained from the Department of Pathology, University of Granada. All samples were fixed in 10% formalin in PBS 0.1 M for 8 to 12 hr and embedded in paraffin with the conventional protocol. Samples were cut 5 µm thick for staining with the conventional HE method and the FM method described in Bancroft and Gamble (2008), as well as contrasted with nuclear Fast Red (ref: AR180 from DakoCytomation, Glostrup, Denmark) and the new histochemical FMPS method.

The FMPS Procedure

Tissue sections were deparaffinized and hydrated with the conventional protocol. Then they were incubated in an FM working solution (see solution 1 below for composition) in a microwave for 3 cycles of 50 sec at maximum power. Each of the successive steps was followed by three washes in distilled water. The sections were differentiated in 2% gold chloride for 5 min (ref: 481130 from Sigma-Aldrich, Steinheim, Germany), and the fixation of silver was carried out with 2% sodium thiosulfate (ref: HT1005; Sigma-Aldrich) for 1 min. The picrosirius staining (see solution 2 below for composition) was performed for 30 min at room temperature (Junqueira et al. 1979). The nuclear contrast was performed with Harris hematoxylin (ref: 253 949, from Panreac, Barcelona, Spain) for 3 min. The sections were dehydrated, cleared in xylol, and mounted in hydrophobic medium.

The controls corresponded to a pigmented sample in which melanin had been previously blanched with 10% hydrogen peroxide (ref: 141 076; Panreac). This control was performed at room temperature, and the reaction was controlled under light microscope. After bleaching, the FMPS method was carried out, as described previously.

The sections were examined under a Nikon Eclipse 90i microscope, and images were captured with a Nikon Digital Camera DXM 1200c and NIS Elements software (Nikon, Tokyo, Japan) for light microscopy. The polarized light microscopy was performed with a Olympus BX 51 microscope, and images were captured with an Olympus digital camera DP70 and DP Manager software (Olympus Optical, Tokyo, Japan).

Solution 1: Fontana-Masson

Stock solution:

• 10% silver nitrate (ref: 131459; Panreac) 95 ml. Add ammonium hydroxide (ref: 131130; Panreac) drop by drop, until the solution precipitates and clears again.

• 10% silver nitrate 1 ml. The solution will become slightly cloudy; allow undisturbed for 4 to 24 hr.

Work solution:

• Stock solution 2.5 ml

• Distilled water 47.5 ml

Solution 2: Picrosirius

• Sirius red F3B (ref: 365548; Sigma-Aldrich) 0.2 g

• Picric acid (ref: 80450; Sigma-Aldrich) saturated solution 100 ml

Quantification of the Reaction and Statistical Analysis

For quantitative analysis, the 52 biopsies were divided into three groups: HE, FM, and FMPS. Points were selected in regions with melanin, collagen fibers, epithelium, and the area without tissue (control) for determining the intensity in RGB (red, green, and blue) channels with the Image J software (National Institutes of Health, Bethesda, MD). The intensity calculation was performed by comparing the RGB intensity of the regions of interest with the control in the three methods used.

We calculated the mean of intensity for the variables—the melanin pigment, the collagen fibers, and the epithelium—in the three histological methods used. Statistical significance was determined using the Student t-test, and all p values below 0.05 were considered statistically significant for the two-tailed test.

We performed a three-dimensional surface color plot to show the intensity of the reactions in specific regions with NIS Elements software (Nikon, Tokyo, Japan).

Results

Morphological Analysis

Hematoxylin eosin

In the sections stained with HE, we observed the cellular and tissue patterns, which allow us to recognize the tissue and to establish the diagnosis based on morphological parameters (Fig. 1A,D,G). We can recognize the presence of a pigment whose color varied from brown to black with a cytoplasmic pattern, except in those cases in which the pigment was abundant and diffuse (Fig. 2A,D).

Figure 1.

Illustrative section of normal skin with a pigmented basal layer (A–C with inserts). Note the positive argentaffin reaction in the basal layer (B) and the morphological characteristics, positive argentaffin reaction, and histochemistry demonstration of collagen fibers in red (C). Section of blue nevi (B. NEVI) with a pigmented melanocytic infiltration in the stroma (D–F with insets). Section of desmoplastic melanoma, poorly pigmented (G–I). Note the bands of collagen fibers in red (I) and cells with a positive argentaffin reaction in the cytoplasm (I insets). HE, hematoxylin-eosin; FM, Fontana-Masson; FMPS, Fontana-Masson picrosirius. Bar = 100 µm.

Figure 2.

Illustrative section of a melanocytic intradermal nevi (MIN) composed of poorly pigmented neoplastic cell nests (A–C). High magnification of MIN, where it is possible to observe the neoplastic cell nests and stromal collagen fibers (D–F). Note the morphological characteristics, positive argentaffin reaction in black, and the histochemistry demonstration of collagen fibers in red (F). MIN surface color plot of the three staining methods (G–I).Bar = 100 µm.

In the stroma, we could identify acidophilus components of the extracellular matrix (ECM) and the cellular elements present. It is not possible to determine the limits of the lesion or to observe changes in the remodeling of the ECM (Fig. 1A,D,G and Fig. 2A,D).

The three-dimensional surface color plot shows the reaction intensity for the different elements and their tissue distribution (Fig. 2G), showing scarce staining reaction for isolated melanocytic clusters, whereas the stroma showed a homogeneous stain intensity.

Fontana-Masson

The argentaffin reaction was observed in the melanosomes, which confirms the presence of the melanin pigment in all samples tested. However, it was not possible to properly observe morphological characteristics (Fig. 1B,E,H). With high magnification, it is possible to recognize the pigment with a cytoplasmic pattern (Fig. 2B,E).

The three-dimensional surface color plot shows the intensity of the reaction for melanin and its distribution (Fig. 2H). It shows an intense argentaffin reaction for the melanin and very little staining for other component.

Fontana-Masson picrosirius

With this method, it was possible to recognize the cell and tissue patterns that allowed us to establish the diagnosis of lesions based on morphological parameters (Fig. 1C,F,I). An argentaffin reaction was observed in black, confirming the presence of the melanin pigment and its distribution in all the samples (Fig. 1C,F,I). With this method, it is possible to observe morphological characteristics and a positive argentaffin reaction in the cytoplasm of the cells (Fig. 2C,F).

The stroma could be clearly distinguished from the epithelium and the melanin pigment. The melanocytic lesions presented an intense histochemical reaction for collagen fibers in red (Fig. 1C,F,I).

The three-dimensional surface color plot showed the intensity of the reaction and the distribution of the melanin pigment and the collagen fibers, easily allowing the identification of collagen fibers, epithelium, and melanin. The three-dimensional surface color plot allowed us to identify the regions with greater intensity of reaction. This software showed that the FMPS method presented a greater overall intensity of reaction when compared to FM and HE in the same regions (Fig. 2G,H,I).

In relation to the stroma, we properly observed the organization, distribution, and fibrillar structure of collagen fibers. In neoplastic lesions, it was possible to recognize the cellular component organization and nuclear atypia. We also observed an intense argentaffin reaction for the melanin pigment and the reorganization of the stromal collagen fibers around the lesion (Fig. 2C,F) but could not identify the margin of the tumor.

With polarized light microscopy, we observed the increase of the birefringence of collagen fibers and their organization and distribution (Fig. 3). Strikingly, this was not observed when HE or FM was used separately (data not shown).

Figure 3.

Fontana-Masson picrosirius (FMPS) stain (A–H). Arterial permeation by pigmented melanoma cells in light microscopy (A). The same case with polarized light microscopy that shows a low birefringence of collagen fibers in the vascular wall (B) (bar = 100 µm). Melanoma composed of neoplastic cell nests with little pigmentation. Collagen septa are seen as intense red in light microscopy (C) and low birefringence of collagen fiber in polarized light microscopy (D) (bar = 100 µm). Section of melanocytic intradermal nevi (MIN) composed of neoplastic cell nests with little pigmentation surrounded by collagen fibers (E) and low birefringence of collagen fibers around the cell nests in polarized light microscopy (F) (bar = 100 µm). Section of blue nevi with pigmented cells in the stroma rich in collagen fibers in light microscopy (G) and intense birefringence of collagen fibers in polarized light microscopy (H) (bar = 50 µm).

Quantitative and Statistical Analysis

Intensity analysis of the melanin pigment in 52 cases studied showed that FMPS had a mean intensity slightly higher than FM, without any statistically significant difference (p > 0.05). However, the HE staining had a significantly lower mean intensity when compared to FMPS (p < 0.05) (Table 1).

Table 1.

Analysis of the Intensity in RGB Channel Quantification for Collagen, Melanin, and Epithelium

Method	Number of Cases	Collagen	Melanin	Epithelium
HE	52	39.70 ± 19.55 p = 0.00000	168.58 ± 50.56 p = 0.00008	68.24 ± 31.45 p = 0.29375
FM	52	9.52 ± 6.34 p = 0.00000	189.32 ± 60.72 p = 0.06837	18.98 ± 13.15 p = 0.00000
FMPS	52	140.52 ± 39.93	209.98 ± 50.19	75.1 ± 33.21

All values are shown as mean ± standard deviation of 52 samples analyzed. Statistical p values for the comparison of HE and FM techniques versus FMPS are shown. HE, hematoxylin-eosin; FM, Fontana-Masson; FMPS, Fontana-Masson picrosirius; RGB, red, green, and blue.

Intensity analysis of the collagen fibers in 52 cases studied revealed that the FMPS method had a significantly higher mean intensity compared with the HE staining. The FM method showed very low intensity levels that were significantly lower than those obtained for HE and FMPS (both with p < 0.05) (Table 1).

The determination of the intensity for the epithelial component in the 52 cases studied showed that FMPS did not present significant differences compared to the HE method. However, the FM method was significantly lower than the FMPS and the HE methods (both with p < 0.05) (Table 1).

Discussion

The observation of melanocytic lesions with HE allows us to evaluate the morphological characteristics and diagnosis in most cases. For cases with a difficult diagnosis, it is necessary to carry out a histochemical method such as FM for the identification of the melanin pigment. Currently, for diagnosis of these lesions, it is necessary to have specific immunohistochemical markers, in addition to the histochemical techniques (Taylor and Cote 2005; Stevens and Lowe 2000).

HE staining is nonspecific for many tissue elements because of their electropolar nature, and it is not possible to specifically identify the present pigments, which allow the diagnosis of several diseases.

The new FMPS method allows us to observe the morphological characteristics and to establish the diagnosis of lesions as well as HE and has the advantage of identifying the melanin pigment specifically with the same intensity and sensitivity as a conventional FM method, which is not possible with the HE method.

In relation to the stroma and the ECM protein, with HE, it is possible to recognize acidophilic elements such as the collagen fibers. However, this staining is nonspecific because these elements are stained with the same tone and intensity. The identification of ECM protein acquires an important role in these cases where its fibrillar elements are altered (Murali et al. 2009; Ferrara et al. 2005; Provenzano et al. 2006). In some cases, it is necessary to evaluate the ECM integrity to establish the diagnosis of an invasion in a variety of neoplasms (Leber and Efferth 2009; Provenzano et al. 2006).

In the blue nevi, Spitz nevi, and related lesions, the degree of dermal sclerosis (the characteristics of collagen bands, the neoplastic cell organization, and pigmentation degree) should be evaluated for establishing the diagnosis (Murali et al. 2009; Ferrara et al. 2005). In these lesions, numerous techniques are currently used and evaluated in parallel.

Melanomas are characterized by their ability to proliferate, their proteolytic activity, their neo-angiogenesis and metastasis, and their capability to remodel several ECM proteins (Leber and Efferth 2009; Srivastava et al. 2003; Rudolph and Matrisian 1998). With the FMPS method, besides the above-described advantages, it is possible to identify with high intensity and specificity the collagen fibers, which is not possible with the HE method.

The identification of collagen fibers is highly specific with the FMPS method because it uses a strong anionic polyazo dye called sirius red. This dye interacts with cationic groups on the surface of the collagen molecules in parallel, giving an intense red color to the fibers in light microscopy and enhancing the natural birefringence of collagen fibers in polarized light microscopy (Junqueira et al. 1979). This phenomenon does not occur with the HE and FM methods because birefringence is particularly induced by picrosirius staining. With the new FMPS method, it is possible to observe the stromal integrity and collagen reorganization at the tumor-stromal interface, but it cannot identify the margin of the tumor. This allows the identification of cell nests, the different degrees of dermal sclerosis, and the integrity of the vascular and lymphatic vessel walls, thus facilitating the identification and evaluation of vascular and lymphatic permeation in these types of lesions.

The argentaffin reaction of the melanin is not affected by subsequent procedures in this method. Instead, the intensity of the reaction slightly increases compared to the conventional FM method. The histochemical reaction for collagen fibers is not affected by treatment with a high temperature of the samples and provides a contrast that allows us to identify the components separately and specifically. It also has the advantage of increasing the natural birefringence of collagen fibers and to selectively identify them using microscopy of polarized light, which has been characterized by other authors (Junqueira et al. 1979; Montes and Junqueira 1991; Trau et al. 1991).

The quantitative and statistical analysis shows that the new histochemical FMPS method is not inferior to the conventional FM method and no less than the HE method for detecting the melanin pigment and observing the morphology, respectively. However, it is significantly superior in the detection of collagen fibers because the intensity is particularly increased by picrosirius staining in light microscopy. Because of the sensitivity and specificity of this new method, it is possible to evaluate morphological patterns of the elements present in the sample, identify the melanin pigment with high intensity and specificity, and assess the state of the ECM with the specific identification of stromal collagen fibers. The three-dimensional surface color plot allows us to identify regions with greater intensity of reaction. This software shows that the new histochemical FMPS method has a high overall intensity of reaction compared to the HE and FM methods in the same tissue and regions.

In conclusion, the new histochemical FMPS method allows us to evaluate histological patterns, the argentaffin reaction of the melanin pigment, and the histochemical reaction of the stromal collagen fibers simultaneously. This method combines the properties of sensitivity and specificity of reagents used, providing a high intensity and specificity of reaction in the identification of the microscopic structures described above. Currently, an available histochemical method with these characteristics does not exist. Furthermore, the new FMPS method is carried out in just 60 min, being a fast, simple, and inexpensive procedure. Although this is not a method that aims to replace the procedures used routinely, it could be useful not only to help in a better diagnosis of pigmented melanocytic lesions but also to evaluate the correlation between normal and neoplastic pigmented cells and the surrounded stromal pattern simultaneously. For all these reasons, we hypothesize that this double staining of collagen fibers and melanin pigment could be useful in pathology.