Evaluation of an enzyme immunoassay and immunodiffusion for detection of anti‐Histoplasma antibodies in serum from cats and dogs

Abstract Background Histoplasma antigen and anti‐Histoplasma antibody detection are used to support the diagnosis of histoplasmosis. There is a paucity of published data on antibody assays. Objectives Our primary hypothesis was that anti‐Histoplasma immunoglobulin G (IgG) antibody detection using enzyme immunoassay (EIA) will be more sensitive as compared to immunodiffusion (ID). Animals Thirty‐seven cats and 22 dogs with proven or probable histoplasmosis; 157 negative control animals. Methods Residual stored sera were tested for anti‐Histoplasma antibodies using EIA and ID. Results of urine antigen EIA were reviewed retrospectively. Diagnostic sensitivity was calculated for all three assays and compared between immunoglobulin G (IgG) EIA and ID. The diagnostic sensitivity of urine antigen EIA and IgG EIA, interpreted in parallel, was reported. Results Sensitivity of IgG EIA was 30/37 (81.1%; 95% confidence interval [CI], 68.5%‐93.4%) in cats and 17/22 (77.3%; 95% CI, 59.8%‐94.8%) in dogs. Diagnostic sensitivity of ID was 0/37 (0%; 95% CI, 0%‐9.5%) in cats and 3/22 (13.6%; 95% CI, 0%‐28.0%) in dogs. Immunoglobulin G EIA was positive in all animals (2 cats and 2 dogs) with histoplasmosis but without detectable antigen in urine. Diagnostic specificity of IgG EIA was 18/19 (94.7%; 95% CI, 74.0%‐99.9%) in cats and 128/138 (92.8%; 95% CI, 87.1%‐96.5%) in dogs. Conclusion and Clinical Importance Antibody detection by EIA can be used to support the diagnosis of histoplasmosis in cats and dogs. Immunodiffusion has an unacceptably low diagnostic sensitivity and is not recommended.


| INTRODUCTION
Histoplasmosis is an enzootic invasive fungal infection (IFI) of mammals worldwide. Veterinary species most commonly infected include domestic cats and dogs. After inhalation of microconidia found in soil, disease can be localized to the respiratory tract or disseminate via blood or lymphatics. After dissemination, multisystemic disease is most common, but disease can be localized to any organ. 1,2 Affected sites include bones, joints, gastrointestinal (GI) tract, eyes, or skin. [1][2][3][4] Diagnosis can be made by finding Histoplasma yeast organisms in tissue or body fluid samples, but doing so is not possible in some cases. Failure to identify the organism could be a result of disease in anatomic locations that preclude safe sampling or samples that contain low numbers of organisms. Culture of Histoplasma from affected tissues or body fluids is also confirmatory, but it is uncommonly used clinically, because turnaround time can be long (≥4 weeks). 5 Because of these limitations, non-culture-based biomarkers often are used.
These include Histoplasma antigen or anti-Histoplasma antibodies. A commercially available enzyme immunoassay (EIA) detecting Histoplasma antigen (MVista Histoplasma Quantitative Antigen EIA, Mira-Vista Diagnostics, Indianapolis, IN) in urine, has a diagnostic sensitivity of 89% to 95% in cats and dogs. [6][7][8][9] Although cross-reactivity occurs with closely related fungal organisms occurs such as Blastomyces, the test remains clinically useful because specificity for fungal vs nonfungal disease is 97% to 100%. [6][7][8] Histoplasma antigen testing also is used for treatment monitoring because concentrations decrease with successful treatment and increase with disease relapse. 10 For the approximately 5% to 10% of cats and dogs with histoplasmosis that do not have detectable Histoplasma antigen in urine, detection of anti-Histoplasma antibodies might be clinically useful.
Currently, commercially available tests are either an enzyme immunoassay (EIA) or immunodiffusion (ID). Multiple diagnostic service laboratories in the United States offer ID for detection of anti-Histoplasma antibodies, some as part of so-called "fungal panels." Immunodiffusion utilizes a clear agarose gel with multiple wells in close proximity. Purified Histoplasma mycelial H and M antigens, a catalase and β-glucosidase, respectively, are used. Both antigens are added to one well and patient serum and positive control serum are added to adjacent wells.

| MATERIALS AND METHODS
Residual sera stored from cats and dogs enrolled in other clinical studies were used for antibody testing. 10,12 All samples were collected in accordance with study protocols approved by the respective Institutional Care and Use Committees. Pet-owner signed consent was obtained at the time of study enrollment. Serum samples were from animals with histoplasmosis at the time of diagnosis and animals without histoplasmosis, either with an alternative diagnosis or healthy control animals.

| Diagnosis and classification of histoplasmosis
Proven histoplasmosis required finding Histoplasma yeast organisms on cytopathology or histopathology by a board-certified veterinary clinical pathologist or board-certified veterinary anatomic pathologist, respectively. Probable histoplasmosis required consistent clinical findings and detectable Histoplasma antigen in urine using enzyme immunoassay (EIA). Disseminated histoplasmosis was defined as clinical evidence (e.g., laboratory test results, pathology, imaging results) of disease in any body system other than lung. Disseminated disease was further defined as disseminated localized or disseminated multisystemic. Disseminated localized disease was defined as disease of a single organ or organ system and associated lymph nodes (LN), other than the lung. Disseminated multisystemic disease was defined as clinical evidence of disease in ≥2 organ systems or finding Histoplasma organisms in blood (fungemia). Pulmonary histoplasmosis was defined as disease apparently localized to the lung and associated intrathoracic LN. Only sera collected from animals within 1 month of diagnosis were used to determine the diagnostic sensitivity of the anti-Histoplasma antibody assays.
Data retrieved from medical record review for animals with histoplasmosis included signalment, clinical signs, and diagnostic test results (Histoplasma antigen EIA on urine, CBC, serum biochemistry, imaging studies, cytopathology and histopathology, and infectious disease and endocrine testing).

| Control animals not diagnosed with histoplasmosis
Negative control sera were convenience samples from animals not diagnosed with histoplasmosis. These were primarily client-owned animals from an area where histoplasmosis was enzootic along with a smaller number of sera from purpose-bred research animals not expected to have been exposed to Histoplasma.  Histoplasma antigens and blocking buffer. Patient sera, standards, and controls were added to wells. After incubation at 37 C for 1 hour, wells were washed, and biotinylated anti-canine or anti-feline IgG was added to each well. The microplate was incubated at 37 C for 1 hour.
The wells were washed, and streptavidin-horse radish peroxidase conjugate was added to each well and then incubated at 37 C for 1 hour.
The wells were washed, and chromogen solution containing peroxidase substrate was added and incubated at room temperature for 10 minutes. The enzymatic reaction was stopped by adding 2 N sulfuric acid. The plate was read at a dual wavelength of 450/620 nm.
Immunoglobulin G antibody concentrations <8.0 enzyme immunoassay units (EU) were considered negative, 8.0 to 9.9 EU were considered indeterminant, and ≥10.0 EU were considered positive.

| Anti-Histoplasma antibody detection in serum using immunodiffusion
Testing was performed in accordance with reagent manufacturer instructions (Fungal Immunodiffusion Reagents, Meridian BioScience, Cincinnati, OH) based on the Ouchterlony double-diffusion method. 15,16 In short, purified Histoplasma mycelial H and M antigens were added until full to a well on clear 0.9% agarose gel. Positive control serum and patient serum were added until full to adjacent wells, respectively. Gels were incubated at room temperature in airtight containers containing a wet gauze to avoid drying. Gels were inspected daily for up to 72 hours. Gels were read against a dark black background with a bright indirect light source. To be considered valid, two visible bands of immunoprecipitation, one for each antigen, must have been present between the control serum and antigen wells. The M antigen band is found closest to the patient serum well, whereas the H antigen band is found closest to the antigen well. To be interpreted as positive, one or two distinct bands must have been present between the patient serum well and the Histoplasma antigen well. In addition, the patient serum band(s) needed to make a smooth junction with the positive serum control band (full identity). If the positive control and patient serum bands crossed (partial identity), or if there was not a distinct visible line of precipitation between patient serum and antigen wells it was considered a negative test result.

| DISCUSSION
Our study provides diagnostic performance data for two commercially available anti-Histoplasma antibody assays for cats and dogs. Most importantly, we showed that anti-Histoplasma antibody detection by ID has unacceptably low diagnostic sensitivity. In addition, we showed that antibody detection by EIA provides an acceptable diagnostic per- Anti-Histoplasma antibody detection by ID has been used for many decades to support the diagnosis of histoplasmosis in cats and dogs. 11 It is offered commercially by several veterinary diagnostic laboratories, often as part of so-called "fungal panels," which include antibody detection by ID for multiple pathogens. Until our study, no published data was available regarding the diagnostic performance of ID for histoplasmosis. Our findings suggest that histoplasmosis should not be ruled out based on a negative ID antibody test. In fact, the diagnostic sensitivity is so low that the use of ID for histoplasmosis in cats and dogs cannot be recommended.
As in veterinary medicine, ID has long been used for anti-Histoplasma antibody detection in human medicine. In one study the diagnostic sensitivity was 37% in people with acute pulmonary histoplasmosis. 13 The sensitivity increased to 56% when people exposed during point-source outbreaks also were included. 13 In a separate study, sensitivity was 78% for people with disseminated or pulmonary histoplasmosis. 17  In humans, concurrent immunosuppressive conditions can lead to a blunted humoral immune response and false negative ID antibody testing. 17 This situation is most often encountered with disseminated histoplasmosis. 17 Although disseminated disease is most common in cats and dogs, comorbidities causing immunosuppression are uncommonly identified and such a situation was considered an unlikely cause of low ID antibody detection in our study. 2,9,18 Immunodiffusion has been used successfully to detect other fungal pathogens. Most notably, it is a common biomarker test used to support the diagnosis of coccidioidomycosis (Valley Fever). 19,20 For coccidioidomycosis, ID has a diagnostic accuracy similar to EIA with a reported sensitivity of 97% to 100% and 87% to 89% in cats and dogs, respectively. [19][20][21][22] Immunodiffusion also has been used to detect anti-Blastomyces and anti-Aspergillus antibodies in cats and dogs, although as for histoplasmosis, ID has a lower diagnostic sensitivity as compared to EIA. 14,23,24 More animals with histoplasmosis had detectable antigen in urine as compared to positive IgG EIA. As such, antigen detection in urine should be the initial non-invasive biomarker considered. The pooled average sensitivity of antigen detection in urine in published studies is 93% and 94%, for dogs and cats, respectively, meaning that 6% to 7% of animals will have false negative results. [6][7][8][9] All four animals in our study with histoplasmosis and no detectable antigen were positive on IgG EIA. Thus, when interpreted in parallel (either test is positive and then diagnostic for histoplasmosis), the combined diagnostic sensitivity was 100%. The finding of Histoplasma antigen concentrations (<0.4 ng/mL) in urine is a clinical challenge, because intuitively those close to the diagnostic cutoff are more likely to be false positive.
Although this situation has not been well studied in veterinary species, it has been reported in humans. 25  Finally, the interpretation of ID testing is subjective, requiring visual inspection to identify a band of immunoprecipitation. All of the above variables notwithstanding, it is unlikely a large enough difference exists in the diagnostic performance of ID testing among veterinary laboratories to recommend its use for anti-Histoplasma antibody detection in cats and dogs.
The second limitation is that negative control animals used to establish diagnostic specificity had inconsistent diagnostic evaluations.
These differences could have led to the inclusion of a small number of animals with occult histoplasmosis. The participating hospital was located in an endemic area had clinicians that were very familiar with histoplasmosis. Even in an endemic area, the overall prevalence of histoplasmosis in the participating hospital patient population during the study period was only 0.14% for dogs and 2.3% for cats. The prevalence of histoplasmosis in the control group would be expected to be even lower because animals in this group had at least 1 hospital visit, and in some cases extensive diagnostic evaluations, without being diagnosed with histoplasmosis. Collectively, these considerations suggest that inclusion of animals with occult histoplasmosis in the control group, if it did occur, was very uncommon. Albeit imperfect, our study provides an estimation of diagnostic specificity in an endemic area that will be useful for the practicing clinician. We decided to include all animals not diagnosed with histoplasmosis in the control group to avoid excluding animals without a definitive diagnosis but having overlapping clinical signs with histoplasmosis, which could have falsely inflated the reported specificity. For example, more extensive medical record reviews of the two control dogs with the highest IgG EIA concentrations could have been excluded because of overlapping clinical findings (Table S1). One dog had diffuse interstitial lung disease, pulmonary hypertension, and was receiving immunomodulatory medications for immune-mediated neutropenia. This dog was also positive for anti-Histoplasma ID antibodies. The second dog had tracheobronchial lymphadenopathy and a lytic vertebral lesion. Both dogs were negative for Histoplasma antigen in urine, but neither was tested for anti-Histoplasma antibodies as part of the diagnostic investigation.
In conclusion, anti-Histoplasma antibody detection by EIA is clinically useful, especially when combined with antigen detection. Immunodiffusion has an unacceptably low diagnostic sensitivity and its use for anti-Histoplasma antibody detection in cats and dogs is not recommended.

ACKNOWLEDGMENT
No funding was received for this study.