Polarization microscopy as a tool for quantitative evaluation of collagen using picrosirius red in different stages of CKD in cats
Abstract
Chronic kidney disease (CKD) is a relevant disease in feline clinic. The tubulointerstitial damage, with collagen deposition and fibrosis, is an important result of this process. The aim of this study was to quantify and correlate the deposition of collagen and severity of interstitial fibrosis (IF) in the kidney from cats in different stages of CKD. Kidney fragments from 10 adult cats with CKD were analyzed and stained by Masson’s trichrome (MT) and Picrosirius red (PSR) for circular polar- ized microscopy. Random quantitative analysis was performed on MT sections to classify the degree of IF, per field area, with and without circular polarization. Statistics correlations were per- formed by Spearman’s (q; p < .05). There was a significant correlation of IF quantification with the area of interstitial collagen deposition by polarized PSR (PSRp) (r 5 .7939, p 5 .0098) and nonpo- larized PSR (PSRn) (r 5 .7781, p 5 .0080). There was a positive correlation of serum creatinine (sCr) at different stages of CKD with PSRp (r 5 .7939, p 5 .0098), PSRn (r 5 .8667, p 5 .0027) and MT (r 5 .7818, p 5 .0117). Correlations between the percentage of quantified area was also positive from PSRp to PSRn (r 5 .9030, p 5 .0009) and PSRp to MT (r 5 .7939, p 5 .0098). The PSRN was also correlated with MT (r 5 .9273, p 5 .0001). The correlation with IF and sCr follows the disease evolution and the quantification of collagen by PSR is an excellent tool for analyzing the disease severity at different stages.
1| INTRODUCTION
Chronic kidney disease (CKD) is one of the main illnesses that culmi- nate in death in cats (Egenvall et al., 2009). Progressive kidney diseases lead to a fibrosis process that deprives the body of its function and promotes consequent decline in kidney function (Eddy, 2000; Hakim & Lazarus, 1989; Prunotto et al., 2011). The process is irreversible, lead- ing inevitably to end-stage renal disease (ESRD) (Chakrabarti, Syme, Brown, & Elliott, 2012; Hewitson, 2012; Yabuki et al., 2010).Chronic kidney disease is common in older cats (Bartges, 2012; Marino et al., 2014.). The age configures, in this sense, as one of thepossible factors involved in the establishment of CKD in this species (Brown, Elliott, Schmiedt, & Brown, 2016). Senility leads to a decrease in the number of nephrons and consequent decrease in organ function as a whole. It is worth noting that the degree of functional impairment does not always reflect the structural loss (Bartges, 2012).Correlating senility with the development of interstitial fibrosis in kidneys from senile cats requires research efforts in a molecular level, to demonstrate their direct contribution to the tubular cell death, inflammation and fibrosis process, that are not yet fully clarified in cats (Brown, Elliott, Schmiedt, & Brown, 2016; Lawler et al. 2006). Tradi- tionally, trichrome stains have been used to detect collagen fiber in tissue sections from kidneys with interstitial fibrosis. However, there is athat can, under certain circumstances, lead to a substantial underesti- mation of the collagen content in the analyzed sample (Kiernan, 2002; Whittaker, Kloner, Boughner, & Pickering, 1994).Polarized light microscopy is a powerful tool to study the fibers orientation in tissues containing fibrillar collagen (Changoor et al., 2011).
The use of picrosirius red stain under polarized light micros- copy for detection and analysis of collagen was first described by a Brazilian group (Junqueira, Bignolas, & Brentani, 1979; Vidal, Mello, & Pimentel, 1982). Despite the specificity conferred by the combina- tion of color and polarized light, the method is not as fully explored as it might be (Rich & Whittaker, 2005; Whittaker & Canham, 1991). One explanation for this under-utilization is the inability to visualize all types of collagen when the polarized light is linearly used, since fibers or their portions appear dark when in parallel alignment to the transmission axis of any polarization filter. In contrast, the use of cir- cularly polarized light eliminates this problem and enables the visual- ization of each fiber portion (Rich & Whittaker, 2005; Whittaker et al., 1994).Fibrotic lesions are considered relevant and present in different stages of chronic kidney disease in cats. Nowadays there is no effective treatment to control or slow the progression of fibrotic process in cats and, consequently, delay the disease progression. Several efforts have been made to provide a better understanding of the renal interstitial fibrosis and control its development process; although fibrotic lesions in feline CKD have been described, there are no studies quantifying the collagen deposition under polarized light and correlating these lesions with the disease stages. The aim of this study was to quantify and correlate the deposition of collagen and severity of interstitial fibrosis in kidneys from cats at different stages of CKD.
2| MATERIAL AND METHODS
The experimental group consisted of 10 cats of both sexes, being 6 males and 4 females, aged between 5 and 15 years (7.9 6 1.2 years) and evaluated in the Veterinary Hospital from Universidade Estadual do Cear´a, Cear´a, Brazil from 2014 to 2015. The inclusion criteria were: history of kidney disease through clinical, laboratory, and imaging eval- uation, with death during that period (not necessarily due to kidney dis- ease), not presenting any evidence of acute kidney disease, neoplasia, and ureter/urethral obstruction.Cats selected in this study were treated animals that died during the study period, having a clinical history compatible with CKD and which have different causes of death, such as pancreatitis, poisoning, among others, which were sent for necropsy.The calculation was performed using the sample calculation for qualitative variable (incidence of CKD 5 20%), considering a finite pop- ulation of animals, evaluated during a period of time, that had consist- ent changes with CKD. The sample size was calculated as follows:n5 N : ðp : qÞ : Z =2 where n: sample size; N: population size; p: proportion of favorable results in the population; q: proportion of unfavorable results in the population; Za/2: Critical value for the confidence level (95% 5 1.96); E: absolute precision (5%).This study was approved by the Ethics Committee for Animal Use of Universidade Estadual do Ceara´, according to the principles of the Bra- zilian College of Animal Experimentation (Memo no. 11585871-7/10). All data were anonymously analyzed and a written consent was signed by the animals’ owners.Both kidneys were collected and measured after the animals’ death.
A small representative sample from the renal cortex of the right kidney, in full cross section of about 1.5 cm2, was collected, fixed in 10% buf- fered formalin and processed by standard histological processing tech- niques. Sections (2 mm) were stained by Masson’s Trichrome (MT) and Picrosirius red (PSR) modified technique by Constantine and Mowry (1968).A random quantitative analysis by semi-quantitative scoring system was performed in 20 random fields by an experienced pathologist (Viana, D A), according to the methods previously established by Raij, Azar, & Keane (1984). Sections stained by MT were analyzed to classify the interstitial fibrosis degree. Lesions were classified by severity score from 0 to 4, where 0 5 no interstitial fibrosis; 1 5 mild (interstitial fibrosis within 25% of the area from the image capturing field); 2 5 moderate (interstitial fibrosis from 26 to 50% field); 3 5 severe (interstitial therefore, the total score quantification of fibrosis is per- formed by the following equation:Total score5 ðn:11n:21n:31n:4Þ20where n represents the number of fields that received the respective score for each slide.This quantification method is supported by quantitation studies for scores (Gibson-Corley, Olivier, & Meyerholz et al., 2013) and quantifi- cation of interstitial fibrosis according to the described by Yabuki et al. (2010) and S´anchez-Lara, Elliott, Syme, Brown, and Haylor (2015). Analyses were performed on 20 random fields (1003) under a trinocu- lar microscope (Nikon® Eclipse Ni with a Nikon® DS RI1 camera attached).The quantitative analysis of cortical renal interstitial fibrosis per field area was performed according to a modified technique by Encarnacion et al. (2004).
The morphometric study of tissue sections stained by MTstained by PSR were analyzed separately under polarized and nonpo- larized light.Random fields (n 5 10) from the cortical area of each slide stained by PSR (1003) were captured, with and without circular polarization (Nikon Eclipse Cipol microscope with an attached Nikon DS-Ri2 digital camera). Digital images of histological sections stained by MT were captured in a standardized way (1003) through a trinocular microscope (Nikon® Eclipse Ni with an attached Nikon® DS RI1camera). Random fields (n 5 10) were captured from the kidney of each animal (1003).Captured images (TIFF format; 1608 3 1608 pixels, corresponding to 11 MB), were used to select regions of interest (ROI), corresponding to tubulointerstitial fibrosis, excluding tubular, vascular, and glomerular epithelium. An initial arithmetic subtraction of the background lighting was performed and images were readjusted to a RGB Stack Model Software. Then these images were adjusted by the threshold tool, and areas related to interstitial fibrosis were delimited, with subsequent quantification of total collagen per area in each analyzed field.The total area of each ROI was calculated and expressed as a per- centage of the total image area. The percentage of the total area quan- tified in 10 random fields for each kidney from each animal was obtained according to the mathematical formula: (R areas quantified%/ 10). All measurements derived from the analysis were automatically transferred to a Microsoft Excel spreadsheet and subjected to statisti- cal analysis.
The total area of each ROI was calculated and expressed as a per- centage of the total image area. The percentage of the total area quan- tified in 10 random fields for each kidney of each animal was obtained according to the mathematical formula: (R areas quantified%/10). All measurements derived from the analysis were automatically transferred to a Microsoft Excel spreadsheet and statistical analysis was performed.The aim was to distinguish the birefringent (bright) from the birefrin- gent (dark) materials. Captured color images were binarized (black and white). Birefringent regions, which correspond to positive staining, were converted to white. A threshold was set in black areas were excluded and the remaining volume of the white area was quantified in relation to the captured field area (Figure 1).The aim was to distinguish positively stained material (red) from unstained one (yellow) and tubular regions (lumen/empty space; white) of the staining within the ROI. Captured color images were binarized (black and white) where the stained material was positively constituted of darker regions. Therefore, the areas to be quantified were adjusted and the percentage quantification of the delimited area was applied, in relation to the total area of the captured field (Figure 1).GraphPad PRISM® v6.0 (GraphPad Software, CA, USA) and Action® v2.7 software were used. Statistical analysis was performed using the Spearman’s rank correlation coefficient and results were expressed by the Spearman’s correlation coefficient (q), considering p < .05.
3| RESULTS
All cats from this study were staged for CKD by serum creatinine (sCr) according to the IRIS staging 1 to 4. In stage 1, animals have sCr values below 1.6 mg/dL, associated with image changes or clinical history sug- gestive of chronicity for kidney disease. One animal was classified at this stage, with sCr 5 1.2 mg/dL and ultrasonographic image indicating loss of corticomedullary definition. Animals belonging to stage 2 (sCr 5 1.6–2.8 mg/dL) had sCr 5 1.8 and 2.1 mg/dL. In stage 3(sCr 5 2.9–5.0 mg/dL) 2 cats were also ranked, with sCr 5 3.0 and4.0 mg/dL and ultrasound images showing irregularities of renal cortex surface. In stage 4, sCr levels exceed 5.0 mg/dL; 5 cats were classified at this stage, with sCr values ranging from 6.0 to 14.0 mg/dL (Table 1).Interstitial fibrosis scores were assigned to kidneys at different stages of CKD (Figure 2). The degree of interstitial fibrosis followed the stage in the samples analyzed, being possible to observe that the fibrosis area, and consequently the deposition of collagen, is larger in more advanced stages.Quantitative analysis by interstitial fibrosis scores in sections stained by MT was correlated with the percentage of the interstitial collagen deposition total area, highlighted by PSRp and PSRn. Total quantification values of each animal is shown in Table 1. There was a significant correlation between semiquantitative quantification by interstitial fibrosis scores with quantification by deposition area of interstitial collagen by PSRp (r 5 .7939, p 5 .0098), and PSRn(r 5 .7781, p 5 .0080) (Figures 3 and 4, respectively).Correlations of sCr quantification with total area of collagen depo- sition in sections stained by PSR under and without polarization were described in the correlation matrix (Figure 3). There was a positive cor- relation of sCr with PSRp (r 5 .7939, p 5 .0098), PSRn (r 5 .8667,p 5 .0027) and MT (r 5 .7818, p 5 .0117). There was a positive corre- lation between PSRp and PSRn (r 5 .9030, p 5 .0009) and PSRP and TM (r 5 .7939, p 5 .0098). The PSRn was also correlated with MT (r 5 .9273, p 5 .0001).
4| DISCUSSION
Interstitial fibrosis has been positively correlated with serum creatinine in cats diagnosed with CKD and staged based on these values accord- ing to IRIS (Chakrabarti et al., 2012; Yabuki et al., 2010). The staging based on the sCr according to IRIS is as an important tool for establish- ing the diagnosis, treatment, and prognosis of CKD in cats (Boyd, Lang- ston, Thompson, Zivin, & Imanishi, 2008; Iris, 2015; Polzin, 2013). Cats evaluated on this study had their sCr measured to be classified in stages as recommended by IRIS. With this step completed, analysis and assignment scores for interstitial fibrosis were initiated, correlating the stage with histopathological findings.
The most common morphologic diagnosis in cats with CKD is chronic tubulointerstitial nephritis and interstitial fibrosis (Chakrabarti et al., 2012; Di Bartola, Rutgers, Zack, & Tarr, 1987; Lawler et al., 2006). Our histological findings showed tubulointerstitial nephritis and degrees of fibrosis with collagen deposition accompanying the stages according to IRIS.In stage 1, interstitial fibrosis was considered from absent to mild and, in the final stages as 3 and 4, scores ranged from severe, in most fields analyzed, to very severe, present even in a few fields observed. These findings corroborate the description by Polzin (2013) and McLeland, Cianciolo, Duncan, and Quimby (2015) where the severity of interstitial fibrosis was significantly higher in advanced stages of CKD when compared to early stages. Interstitial fibrosis leads to an increased production of extracellular matrix and collagen deposition in areas that have suffered injury and subsequent repair (Farris et al., 2011; Hewitson, 2009; Hewitson, Darby, Bisucci, Jones, & Becker, 1998; Lawson, Elliott, Wheeler-Jones, Syme, & Jepson, 2015; S´anchez- Lara et al. 2015; Wynn, 2010).
There was a positive correlation between sCr and quantification per total area of collagen deposition in interstitial fibrosis regions, in sections stained by PSRp (r 5 .7939, p 5 .0098) and PSRn (r 5 .8667, p 5 .0027). This indicates that the degree of tissue damage, resulting in repair (fibrosis), follows the stages, and severity of the disease. Similar results were also demonstrated by Encarnacion et al. (2004) when cor- relating the quantification of fibrosis area in digital images in CKD from transplanted human biopsies.
The correlation of sCr with PSR also showed good accuracy of the method to detect deposition and consequent amount of collagen in areas where fibers are thin. Constantine and Mowry (1968) reported that sections stained by PSR under polarization allowed to distinguish and highlight thin fibers from collagen deposition in fibrotic regions, as well as to increase fiber refrangibility in polarization. Thus, polarization of sections stained by PSR enhanced the detection of collagen fibers not visualized when the PSR is not polarized, increasing the visualiza- tion definition of areas with collagen deposition.sCr values correlated with MT was also positive (r 5 .7818, p 5 .0117), although the best correlation has been in PSRn sections (r 5 .8667, p 5 .0027). This is justified by the fact that PSR, under polarization, defined the deposited collagen more precisely in the stained area. The PSR staining technique used in this study involves differentiation in two colors, where the fibrotic interstitium is clearly marked in red and the other structures in a pale yellow. Trichrome staining has been used for visualization of interstitial fibrosis in routine histological analysis to highlight the fibrotic area in different colors from other structures (Montes & Jun- queira, 1991).
Although MT usually intensely stain collagen fibers, other areas like renal tubules and structures containing collagen, such as reticular fibers and the basal membrane, are not selectively stained, not allowing the differentiation of these structures during quantification, particularly when there are very thin collagen fibers, which can have different col- ors in the same tissue section (Constantine & Mowry, 1968; Montes & Junqueira, 1991; Street et al., 2014; Taboga & Vidal, 2003). Another aspect to be emphasized is that MT might result in a limited method for the quantification of fibrosis in the area, since the differentiation of fibrosis colored blue with a background dyed stained red often becomes a challenge to distinguish pure colors observed (Street et al., 2014).PSR staining has been used as a specific method for collagen detection in tissue sections (Junqueira et al., 1979). In renal tissue, PSR has been used to quantify fibrotic lesions and tubulointerstitial fibrosis (Moreso et al., 1994). Sirius red F3BA was first used as a histological technique in 1964 as a substitute for fuchsin in the Van Gieson Method (Sweat, Puchtler, & Rosenthal, 1964). The tynctorial affinity of Sirius red prepared with picric acid allows very thin fibers to appear bright yellow when viewed under polarization (Puchtler, Waldrop, & Valentine, 1973; Wolman, 1975). The intensity of birefringence by Sir- ius red is superior when compared to other staining techniques to col- lagen, since its molecule binding within the upper groove of collagen types I and III fibrils increases its natural birefringence. When viewed under contrast polarization, collagen is bright on a dark background, thus promoting a better quantification (Junqueira, Cossermelli, & Bren- tani, 1978).
The positive correlation between the PSRp and PSRn (r 5 .9030, p 5 .0009) demonstrated that the polarization confirmed the demarca- tion of collagen area, visualized in nonpolarized sections. Polarization favored the best collagen deposition visualization in areas correspond- ing to interstitial fibrosis, thus allowing quantification per area with better detailing. The quantitation per area allows to correlate the severity of tubulointerstitial damage and repair with the degree of injury and, consequently, with the severity and stage of the disease. It is noteworthy that, in clinical practice, diseases with fibrotic pro- gress are a major challenge since effective antifibrotic agents are not yet widely studied and also little used in cats (Razzaque & Taguchi, 2002).In the face of our results it is possible to conclude that correlation between interstitial fibrosis and sCr follows the disease evolution. Quantification of collagen can be an excellent tool for analyzing the severity of interstitial tubule damage at different stages of CKD and PSR is a good option in this process. Our results validate the use of PSR in the nonpolarized and polarized light microscopy for this purpose and, additionally, emphasize the importance of evaluating a larger number of DIRECT RED 80 samples.