Cryptococcal Glucoxylomannan Does Not Exhibit Cross-Reactivity in the MVista Histoplasma Antigen Enzyme Immunoassay

Abstract

The potential for cross-reaction between Cryptococcus neoformans and Histoplasma capsulatum in antigen assays was evaluated. We tested patient samples, spleens from infected mice, and purified polysaccharides in the MVista Histoplasma antigen enzyme immunoassay and cryptococcal antigen latex agglutination system for cross-reactivity, and none was observed.

The diagnosis of invasive or systemic mycosis is often accomplished through the detection of polysaccharide antigens in body fluids by immunologic assays. The specific polysaccharides detected are galactomannans (GalM) of Aspergillus spp. (2) and Histoplasma capsulatum (1) and glucuronoxylomannan (GXM) of Cryptococcus neoformans (2). We have observed several cases in which tests for both cryptococcal GXM and Histoplasma GalM were positive and where dual infections were documented. However, the literature on dual infection is not extensive (3). This prompted us to investigate further the presence of cross-reactivity between cryptococcal GXM and Histoplasma GalM.

Cryptococcal antigen latex agglutination system (CALAS) kits were purchased from Meridian Bioscience (Cincinnati, OH) and used according to the manufacturer's instructions. The serum samples were treated with pronase before being tested with a CALAS kit. The MVista Histoplasma antigen enzyme immunoassay (EIA) was performed at MiraVista Diagnostics (Indianapolis, IN) as previously reported (1).

Histoplasma and Cryptococcus polysaccharide antigens were prepared and purified as previously described (1, 2). Histoplasma GalM and cryptococcal GXM, galactoxylomannan, and capsular polysaccharide were tested at a concentration of 1 μg/ml, and the CALAS kit positive control was tested undiluted. Cryptococcal GXM, galactoxylomannan, and capsular polysaccharide and the CALAS control were strongly positive (agglutination score of 3+ or 4+) by CALAS assay but negative in the Histoplasma antigen EIA. Conversely, Histoplasma GalM was positive in the Histoplasma antigen EIA (>39 ng/ml) but negative by CALAS assay.

Twenty-nine residual serum specimens from 15 patients with progressive disseminated histoplasmosis were tested by CALAS assay, in a study approved by the institutional review board of Clarian Hospital, Indianapolis, IN. All had underlying immunosuppression, including AIDS in 10, organ transplantation in 2, and miscellaneous causes in 3 patients. The basis for diagnosis of histoplasmosis was Histoplasma antigenemia and antigenuria in seven cases, culture in six cases, and histopathology in two cases. Histoplasma antigenemia ranged from <0.6 to >39 ng/ml (median, 2.5 ng/ml). No samples were positive by CALAS assay (Table 1).

TABLE 1.

Cross-reactivity in clinical specimens and experimental infection

Sample type	No. of positive samples/total no. of samples (% positive, 95% confidence interval)
	Histoplasma antigen EIA	CALAS assay
Histoplasmosis samples
Human serum	29/29 (100, 88.3-100)	0/29 (0, 0-11.7)b
Murine spleena	8/8 (100, 67.6-100)	0/8 (0, 0-32.4)b
Cryptococcosis samples
Human serum or cerebrospinal fluid	0/25 (0, 0-13.3)	24/25 (96, 80.5-99.3)b
Murine spleena	0/12 (0, 0-24.2)	12/12 (100, 75.8-100)b

^aSpleen tissues from mice infected with H. capsulatum were tested in the Histoplasma antigen EIA at a 1:10 dilution and were used undiluted in the CALAS assay. Spleen tissues of mice with cryptococcosis were tested undiluted.

^bP < 0.001 for comparison of Histoplasma antigen EIA and CALAS assay, using Fisher's exact test.

Residual serum or cerebrospinal fluid samples from 25 patients with cryptococcosis were previously described (2). The CALAS assay was positive for 24 of 25 patients, at titers of 1:1 to 1:65,536 (median, 1:1,024). Culture was the basis for diagnosis for the 25th patient. None of the samples were positive in the Histoplasma antigen EIA (Table 1).

In a previously described murine histoplasmosis model (4), eight B6C3F1 mice were infected intranasally with 1 × 10⁶ H. capsulatum organisms and euthanized on day 10 of infection according to institutional guidelines (4). Spleen tissue was homogenized in 2 ml of RPMI medium and tested in the Histoplasma antigen EIA following 10-fold dilution but used undiluted in the CALAS assay. Spleen tissues contained high levels of Histoplasma antigen but were negative by CALAS assay (Table 1).

In a cryptococcal experimental infection model (2), 12 BALB/c mice were infected intravenously with 1 × 10⁴ C. neoformans strain 24067 organisms and then euthanized on day 7 of infection. The spleen tissue was homogenized in 2 ml of RPMI medium and tested undiluted in the Histoplasma antigen EIA and the CALAS assay. Spleen tissues were positive by CALAS assay but negative in the Histoplasma antigen EIA.

In conclusion, no cross-reactivity between Histoplasma GalM and cryptococcal polysaccharides was observed. When both the CALAS assay and MVista Histoplasma antigen EIA are positive, dual infection should be suspected. Note that these findings should not be extrapolated to other Histoplasma or cryptococcal antigen assays.