Sensitivity and Specificity of In Situ Hybridization for Diagnosis of Cutaneous Infection by Leishmania infantum in Dogs

Abstract

An accurate diagnosis of infection by Leishmania infantum in dogs is fundamental for the control of zoonotic visceral leishmaniasis (VL). Histopathology (HP) and immunohistochemistry (IHC) are frequently used for the histological diagnosis of L. infantum in dogs but have shown limited accuracy. To improve the sensitivity and specificity of the histological diagnosis of VL, we evaluated automated in situ hybridization (ISH) using a generic probe for Leishmania and a specific probe for L. infantum in surgical skin biopsy specimens of dogs. The ISH results were compared with those of HP and IHC, using parasitological culture as the reference standard. Skin samples from 51 dogs with cutaneous L. infantum infection and 51 noninfected dogs were randomly selected from samples of dogs from various cities in Brazil where canine VL is endemic. These samples were processed for parasitological culture, HP, IHC, and ISH using both probes. The sensitivities of ISH using the specific probe, ISH using the generic probe, IHC, and HP were, respectively, 74.5%, 70.6%, 69.5%, and 57.6%. The specificity of both ISH probes tested was 100%, and there was no cross-hybridization of the generic and specific probes with selected pathogenic fungi and protozoa. The specific probe discriminated L. infantum from the other species of Leishmania that infect dogs in the New World. ISH is highly sensitive and specific for the diagnosis of L. infantum in histologic samples of skin from infected dogs and can be used on routine biopsy material to make a diagnosis of leishmaniasis.

INTRODUCTION

Leishmaniasis is a worldwide disease caused by protozoa of the genus Leishmania which infect wild and domestic mammals, including humans (1). The spectrum of clinical forms of leishmaniasis can vary from focal cutaneous to disseminated visceral disease (1). There are 12 Leishmania species infecting dogs, L. donovani, L. infantum (syn. L. chagasi), L. major, L. arabica, and L. tropica in the Old World and L. infantum, L. colombiensis, L. panamensis, L. mexicana, L. braziliensis, L. peruviana, L. pifanoi, and L. amazonensis in the New World (1–7). Although many species of Leishmania infect dogs, dogs are considered a proven reservoir only for L. infantum (1, 2). The species L. infantum can also infect humans, cats, and wild mammals and is the cause of zoonotic visceral leishmaniasis (VL) (1, 2, 8).

In many countries, zoonotic VL constitutes a significant public health problem, especially due to its prevalence, high mortality rates, mainly in children, and emergency rate in patients infected by the human immunodeficiency virus (1, 9). Transmission in areas of endemicity is usually via bites of infected sand flies, with dogs as the main domestic reservoir of the parasite (1, 9). Therefore, a rapid and accurate diagnosis of the infection of dogs with L. infantum is fundamental for the control of zoonotic VL transmission.

The clinical diagnosis of canine VL is difficult, with many animals being asymptomatic. Therefore, a variety of serological, parasitological, and molecular laboratory methods have been developed to detect infection by L. infantum in dogs (10, 11). Serological assays, PCR, and parasitological culture are the most sensitive methods for the diagnosis of L. infantum infection (12–18). However, these three methods do not allow visualization of the intact amastigotes within the tissue and correlation of the parasites with associated lesions, which is possible by histopathology (HP) and immunohistochemistry (IHC) (19). HP and IHC are frequently used in the current routine for the histological diagnosis of L. infantum in dogs, but these methods have limited accuracy and do not allow species identification (16, 18, 20). The identification of species of Leishmania is currently only possible by parasitological culture followed by multilocus enzyme electrophoresis (MLEE), which is the reference method, and by PCR (1). Hence, alternative histological methods are necessary to improve the accuracy of diagnosing infection of dogs with L. infantum.

A recently established chromogenic in situ hybridization (ISH) technique is a promising method for the diagnosis of canine VL because it permits the highly specific identification of Leishmania in formalin-fixed, paraffin-embedded (FFPE) surgical biopsy specimens of dogs (19). However, the previously published probe for ISH was unable to determine the Leishmania species (19). In addition, the published ISH protocol was based on manual labeling, which presents lower efficiency and productivity than automation (21). Thus, the present study aimed to evaluate the sensitivity and specificity of automated ISH for the diagnosis of canine cutaneous infection caused by L. infantum in routinely processed surgical biopsy samples using the previously published generic probe for Leishmania and a newly designed probe specific for L. infantum. The results of both ISH tests were compared to those of IHC and HP in the same samples, using parasitological culture as the reference standard technique.

MATERIALS AND METHODS

Samples.

A prospective study was designed using randomly selected samples from 2,066 surgical skin biopsy specimens of dogs collected between the years 2008 and 2012. The dogs originated from seven cities in Brazil with endemic canine VL: Niterói, Rio de Janeiro; Rio de Janeiro, Rio de Janeiro; Bauru, São Paulo; Brasília, Distrito Federal; Cuiabá, Mato Grosso; Palmas, Tocantins; and Fortaleza, Ceará. Skin samples were selected for this study since they are easy to obtain and have been shown to be a good target for the confirmation of canine VL by parasitological culture (14).

For the collection of samples, one 3-mm punch biopsy specimen was obtained from the intact skin over the scapula after disinfection with 70% alcohol and local anesthesia with 2% lidocaine. Each specimen obtained was divided into two samples. One of them was immersed in sterile saline with antimicrobials (22) and submitted for parasitological culture. The other fragment was fixed in 10% neutral buffered formalin and processed for routine paraffin embedding (23). The paraffin blocks were processed for ISH, IHC, and HP. The IHC, HP, and parasitological culture with identification of the species of trypanosomatids by MLEE were performed at the Evandro Chagas Clinical Research Institute, FIOCRUZ, Brazil. The ISH was performed at the Diagnostic Center for Population and Animal Health, Michigan State University. The biopsy procedure performed on animals was approved by the Ethics Committee on the Use of Animals, FIOCRUZ, Brazil (license L-038/08).

Study design.

ISH using a previously published generic (ISH-GP) and a newly developed specific (ISH-SP) oligoprobe for the diagnosis of L. infantum infection in dogs was evaluated. In order to calculate the number of samples required for this study, the estimated values for the ISH were 70% sensitivity/specificity according to preliminary tests, 17% absolute error in sensitivity/specificity, and 5% alpha. Considering the loss of samples during processing, the calculated number was increased by 4%. As a result, 51 dogs positive for L. infantum infection in the parasitological culture and 51 dogs negative for L. infantum infection in the parasitological culture were randomly selected from 2,066 surgical skin biopsy specimens of dogs. Ten of the 51 samples of the L. infantum-negative group were from dogs positive for the protozoon Trypanosoma caninum based on parasitological culture. Thirty FFPE samples (8 positive and 22 negative for L. infantum based on parasitological culture) were consumed after the testing with ISH-GP. They were replaced by new randomly selected samples (8 positive and 22 negative for L. infantum based on parasitological culture) for the testing with ISH-SP. All 132 samples were tested by IHC and HP and a total of 102 dogs was tested with ISH-GP as well as with ISH-SP. Parasitological culture was used as the reference standard to evaluate the sensitivity and specificity of ISH-GP, ISH-SP, IHC, and HP. The microscopic examination of the ISH, IHC, and HP stained slides was performed blindly by a single pathologist with experience in microscopic diagnosis of Leishmania (R.C.M.).

Parasitological culture and characterization by MLEE.

Skin samples collected in saline were seeded in the biphasic culture medium NNN (Novy, MacNeal and Nicolle)-Schneider's insect medium (Sigma-Aldrich Co., St. Louis, MO) containing 10% fetal bovine serum and were incubated at 26 to 28°C. The Leishmania promastigotes isolated were identified by MLEE using five enzymatic systems (24).

Immunohistochemistry and histopathology.

For immunohistochemistry, serial sections of 5 μm were obtained on silane-treated slides and processed according to a previously described protocol (20), with some modifications. The antigen exposure was performed by incubation of the sections in a sodium citrate buffer (pH = 6.0) at 100°C for 20 min in steam. Then, the sections were incubated with rabbit anti-Leishmania polyclonal serum at a dilution of 1:500. Histological sections with numerous Leishmania amastigote forms were incubated with homologous nonimmune serum as the negative control and with the rabbit anti-Leishmania polyclonal serum as the positive control. For histopathology, serial sections of 5 μm were stained by hematoxylin-eosin (23).

Probe design for ISH.

The generic probe is a digoxigenin-labeled oligonucleotide probe that detects a 5.8S rRNA sequence specific to all relevant Leishmania species (19). The specific probe was developed based on previous published sequences from GenBank using the computer program Oligo 6 (25) and following previously described parameters (19, 26). It is an oligonucleotide probe (5′-GCCCCTACCCGGAGGACCAGAAAAGTT-3′) labeled with digoxigenin at the 5′ end (Integrated DNA Technologies, Coralville, IA) that targets a fragment of the kinetoplast minicircle DNA (kDNA) gene. The specific probe was designed to discriminate L. infantum from the other species of Leishmania that infect dogs in the New World, such as L. colombiensis, L. panamensis, L. mexicana, L. braziliensis, L. peruviana, L. pifanoi, and L. amazonensis (3). The in silico analysis using the Basic Local Alignment Search Tool (www.ncbi.nlm.nih.gov/blast.cgi) showed that the specific probe cross-reacted only with L. donovani, L. tropica, and L. major, which are species that do not occur in the New World. Following labeling with digoxigenin at the 5′ end, both oligoprobes were purified by high-performance liquid chromatography (HPLC) (Integrated DNA Technologies, Coralville, IA).

ISH technique.

Several preliminary tests were done to define the best protocol and concentration of each probe in order to achieve the best signal-to-noise ratio and, thus, the maximum sensitivity and specificity. Considering that each probe had different properties of melting temperature at the NaCl concentration of 50 mM, of guanine-cytosine (GC) content, and of molecular weight, as well as different targets and specificities, a different protocol and concentration were used for each probe.

Serial tissue sections of the selected FFPE skin samples were prepared at 5-μm thickness and placed on positively charged slides. These slides were then submitted to deparaffinization and fixation using the Discovery XT automated slide processing system (Ventana Medical Systems, Inc., Tucson, AZ), as programed in the protocol for the RiboMap in situ hybridization reagent system (Ventana Medical Systems). Proteolytic treatment was performed using Protease 3 (0.02 units/ml alkaline protease; Ventana Medical Systems) for 12 min at 37°C. Thereafter, the slides received pretreatment through mild cell conditioning using the citrate buffer-based RiboCC reagent (Ventana Medical Systems) for 4 min at 95°C. The slides were then submitted to denaturation for 4 min at 37°C, followed by hybridization with the antisense oligonucleotide probe for Leishmania suspended in hybridization buffer (RiboHybe; Ventana Medical Systems). The time of hybridization was 1 h at 37°C for the generic probe and at 47°C for the specific probe. The concentration used for the generic probe was 93 ng/ml (1:10,000 dilution) and for the specific probe was 893 ng/ml (dilution 1:1,000). For the generic probe, three stringency washing steps were performed using 0.5× RiboWash (equivalent to 0.5× saline sodium citrate; Ventana Medical Systems) each for 4 min at 42°C. For the specific probe, four stringency washing steps were performed at 42°C using 0.1× RiboWash (equivalent to 0.1× saline sodium citrate; Ventana Medical Systems) for 4 min for the first three and for 8 min for the fourth washing step. After the stringency washes, the slides were incubated with antidigoxigenin antibody for 32 min at 37°C. The antidigoxigenin antibody for the generic probe was a rabbit polyclonal serum (Sigma-Aldrich Co., St. Louis, MO) at a dilution of 1:20,000. For the specific probe, a rabbit monoclonal antidigoxigenin antibody (Invitrogen Corporation, Frederick, MD) was used at a dilution of 1:10,000. After streptavidin-alkaline phosphatase conjugate (UMap anti-Rb AP; Ventana Medical Systems) incubation for 16 min at 37°C, the signal was detected automatically using the BlueMap nitroblue tetrazolium-BCIP (5-bromo-4-chloro-3-indolyl phosphate) substrate kit (Ventana Medical Systems) for 2 h at 37°C. Finally, the sections were counterstained with the nuclear fast red-equivalent reagent Red Counterstain II (Ventana Medical Systems) for 4 min before coverslipping. Sections of FFPE skin and lymph node samples of dogs infected with numerous L. infantum amastigote forms were used as controls. Infection by L. infantum in these controls had been confirmed by parasitological culture and MLEE. For the reagent negative controls, sections were treated only with RiboHybe hybridization buffer. The total duration of slide processing for ISH-GP was 8:17 h and for ISH-SP was 8:25 h using the Discovery XT.

Probe validation.

Before testing the ISH on the selected canine skin samples, validation of each probe was performed to confirm the specificity of the generic probe to Leishmania and the specificity of the specific probe to L. infantum by excluding cross-hybridization with selected pathogenic fungi and protozoa. For this purpose, FFPE tissue samples infected with various protozoa and fungi and FFPE pellets of some of these microorganisms obtained by centrifugation of cultured pathogens were tested. Validation samples of protozoa consisted of L. infantum (skin of dog and pellet of promastigote forms), L. braziliensis (skin of dog and hamster and pellet of promastigote forms), L. amazonensis (skin of mouse and pellet of promastigote forms), Trypanosoma caninum (pellet of epimastigote, spheromastigote, and trypomastigote forms), Trypanosoma cruzi (heart of mouse and pellet of epimastigote and trypomastigote forms), Neospora caninum (lung of rat and pellet of tachyzoites), Toxoplasma gondii (lung and heart of dog, brain of mouse, and intestine and lymph nodes of wallaby), Sarcocystis neurona (brain of horse), and Rangelia vitalii (heart and kidney of dog). The methods used for the diagnosis of protozoa were isolation in parasitological culture and MLEE, except for T. gondii and N. caninum, which were identified by PCR and IHC, and R. vitalii, which was detected by HP in tissues of a dog with the characteristic clinical and pathological alterations caused by this protozoon (27). Validation samples of fungi consisted of Blastomyces dermatitidis (lung of dog), Cryptococcus neoformans (nasal mucosa of horse), Sporothrix (skin of dog and cat), and Histoplasma capsulatum (spleen and kidney of dog and pellet of mycelial form). The diagnosis of all these fungi was based on mycological culture and Grocott's methenamine silver stain. In all pellets and tissue samples used for validation, microorganisms were easily visible by light microscopy.

Statistical analysis.

Data obtained were stored in the EpiData software and then analyzed using the Statistical Package for Social Sciences software (version 16.0) for Windows. The sensitivity and the respective 95% confidence interval (95% CI), the specificity, and the accuracy of ISH-GP, ISH-SP, IHC, and HP were compared to the reference standard (parasitological culture). The comparisons between the sensitivity and specificity of ISH-GP, ISH-SP, IHC, and HP were descriptive.

RESULTS

The ISH using both the generic (ISH-GP) and specific (ISH-SP) probes clearly detected amastigote forms of Leishmania with a dark blue signal that was slightly stronger for the ISH-GP (Fig. 1A). The ISH-SP instead showed less background and a better signal-to-noise ratio (Fig. 1B). There was no cross-hybridization of either probe with any of the other microorganisms tested. However, there was cross-reaction of the polyclonal antidigoxigenin antibody (Sigma-Aldrich) used for ISH-GP with cysts and tachyzoites of Toxoplasma gondii. This problem was solved by replacing this antibody with the same monoclonal antidigoxigenin antibody (Invitrogen) used for ISH-SP. All species of Leishmania tested were detected by ISH-GP (Fig. 1A and D). The ISH-SP detected only L. infantum (Fig. 1B, E, and F).

Fig 1.

(A, B) In situ hybridization on dog skin samples, showing numerous intrahistiocytic L. infantum amastigotes labeled with a dark blue signal using the generic probe for Leishmania spp. (A) and the specific probe for L. infantum (B). Note that the signal for the specific probe is slightly weaker than that of the generic probe and there is less background, which allowed clearer visualization of individualized amastigote forms. (C) Immunohistochemistry of the skin of a dog showing numerous dark-brown-stained L. infantum amastigote forms within macrophages. (D to F) In situ hybridization using the generic probe for Leishmania spp. on the skin of a dog, showing numerous intrahistiocytic L. braziliensis amastigote forms labeled with a dark blue signal (D), the specific probe for L. infantum on the skin of a dog infected by L. braziliensis, demonstrating that amastigote forms (arrows) were not labeled (E), and the same specific probe on the skin of a mouse experimentally infected with L. amazonensis, showing numerous intrahistiocytic amastigote forms that are not labeled (F). Bar = 33 μm in all panels.

The IHC clearly detected amastigote forms of Leishmania with a dark brown signal and a good signal-to-noise ratio (Fig. 1C).

The sensitivity and accuracy results of ISH-SP, ISH-GP, IHC, and HP for detecting L. infantum are listed in Table 1. The specificity of ISH-SP, ISH-GP, IHC, and HP was 100%.

Table 1.

Sensitivity and accuracy of in situ hybridization, immunohistochemistry, and histopathology for detecting L. infantum in skin biopsy samples of dogs

Method (no. of skin samples tested)a	% Sensitivity (95% CI)	% Accuracy
ISH-SP (102)	74.5 (66.1–83.0)	87.2
ISH-GP (102)	70.6 (61.7–79.4)	85.3
IHC (132)	69.5 (61.1–77.9)	86.4
HP (132)	57.6 (49.2–66.1)	81.1

^aFormalin-fixed, paraffin-embedded surgical skin biopsy samples of dogs were tested.

ISH-SP, in situ hybridization using a probe specific for Leishmania infantum; ISH-GP, in situ hybridization using a generic probe for Leishmania; IHC, immunohistochemistry; HP, histopathology.

Of the 51 skin samples positive for L. infantum by parasitological culture, 36 were detected by ISH-GP. Fifteen skin samples positive by parasitological culture were not detected by ISH-GP.

Of the 51 skin samples positive for L. infantum by parasitological culture, 38 were detected by ISH-SP. Thirteen skin samples positive by parasitological culture were not detected by ISH-SP.

DISCUSSION

The ISH-GP and ISH-SP were both sensitive and specific methods for the diagnosis of Leishmania in dogs, showing satisfactory accuracy compared to the reference standard. The values of sensitivity of both ISH methods were very close to those of IHC, with ISH-SP being the most sensitive, followed by ISH-GP, IHC, and HP. The somewhat higher sensitivity of ISH-SP in spite of the slightly lower signal intensity compared to that of ISH-GP may be due to the better signal-to-noise ratio of ISH-SP. The use of ISH and IHC increased the number of correctly diagnosed positive cases in comparison to the results for HP similarly to what has been reported by other authors when using IHC (16, 18, 28, 29).

The results of the current study confirm ISH as an accurate method for the diagnosis of L. infantum infection in dogs. Nonetheless, it is less sensitive for the diagnosis of L. infantum infection in dogs than parasitological culture, PCR on frozen skin samples, and some serological assays (12, 14–18). ISH failed to detect between 25.5% and 29.4% of the parasitological-culture-positive cases in the current study. In the previous study using the same generic ISH probe (19), ISH was negative in 3 dogs out of 6 positive for L. infantum by PCR. In a systematic review (17), the majority of studies on serological assays for the diagnosis of L. infantum infection in dogs found sensitivities higher than 75%. Nonetheless, parasitological culture, PCR, and serological assays present disadvantages that prevent their use alone for the routine diagnosis of L. infantum infection in dogs. The parasitological culture is time consuming, taking from 5 to 30 days (on average, 15 days) to be completed, and there are only a small number of reference centers worldwide currently using MLEE (1, 22). In addition, this method is susceptible to microbiologic contamination, which in many cases prevents its use in samples collected in the field, where proper storage and sterile conditions may be difficult to attain (1, 10). Furthermore, parasitological culture may be difficult to perform due to poor adaptation of some isolates to the medium (1, 10). The drawbacks of PCR are lack of standardization of the different protocols used among laboratories, possibility of contamination, and the fact that it does not necessarily indicate infection with live Leishmania (1, 13, 18, 30). Serologic assays may yield false-positive results due to cross-reactivity with sera of dogs infected with L. braziliensis, T. cruzi, T. caninum, and Ehrlichia canis (15, 31–33). Also, they do not necessarily indicate current infection (10) and do not differentiate positive results produced by natural infection from those induced by vaccines (17).

The main advantage of ISH over parasitological culture, PCR, serological assays, IHC, and HP for the diagnosis of Leishmania infection in dogs is that it simultaneously allows visualization of the intact amastigotes within the tissue and species identification of them as Leishmania infantum, as demonstrated in the present study by ISH-SP. This observation of amastigotes within the tissue, which is not possible by parasitological culture, PCR, and serological assays, offers the possibility to correlate parasites with the associated lesions and, also, to semiquantify the parasite load (16, 19, 34). Similarly to ISH, IHC and HP also link amastigotes of Leishmania to lesions (16); however, they are not able to discriminate L. infantum from other Leishmania species. An important advantage of ISH compared to IHC and HP is therefore the higher specificity of ISH. Extensive testing of the specificity of ISH has shown no cross-reaction of the Leishmania probes with other histomorphologically similar organisms (19), which was confirmed in the current study. Although IHC in the present study was specific for the diagnosis of Leishmania, cross-reactivity with histomorphologically similar fungi, such as Histoplasma capsulatum, has been demonstrated (20). In addition, parasitic organisms such as Histoplasma and Trypanosoma, are difficult to differentiate from Leishmania by HP; thus, wrong or inconclusive etiologic diagnoses may occur using this method (19, 35). Furthermore, commercially available anti-Leishmania antibodies for IHC that work well on FFPE tissues are currently lacking (19, 35). The possibility of automation is another advantage of ISH (21), which was tested with success in the present work but has not yet been evaluated for IHC in the diagnosis of L. infantum infection in dogs. This automation reduced the time to 1 day for the labeling of the slides in comparison to the 2 days needed for the manual protocols of the IHC used in the present study and of the previously reported ISH (19). In addition to productivity, automation will improve the reproducibility of ISH.

The use of ISH-SP for the specific detection of L. infantum will be particularly important in some areas in southeastern Brazil where this protozoal organism co-occurs with L. braziliensis (15, 36). In these areas, many dogs affected only by L. braziliensis are unnecessarily euthanized as a method for controlling VL (9), due to serological cross-reaction with L. infantum. Dogs parasitized by species of Leishmania other than L. infantum do not have to be euthanized in Brazil because the dog is not a proven reservoir of the other Leishmania species and, thus, is not considered to be involved in their zoonotic transmission (9). Considering that the specific probe cross-reacted in silico with L. donovani, L. tropica, and L. major, it is not as useful in areas where these species occur in dogs, such as in the Middle East, Africa, and Asia (4–6, 37, 38), possibly warranting the design of additional probes.

The current study demonstrated that ISH-GP and ISH-SP have a high sensitivity and specificity, improving the histological diagnosis of L. infantum in routinely processed, formalin-fixed surgical skin biopsy specimens of dogs. Hence, the concurrent use of ISH-GP and ISH-SP should be implemented in the laboratory as a useful tool to not only detect Leishmania in general but differentiate L. infantum in surgical samples.