Methods

C. neoformans.C. neoformans strains H99 (C. neoformans var. grubii) and 24067 (C. neoformans var. neoformans) were obtained from J. Perfect (Duke University) and the American Type Culture Collection, respectively. The cells were grown as described elsewhere [12]

C. neoformans–specific MAbs and radiolabeling of MAbsGlucuronoxylomannan-binding murine IgG1 MAb 18B7 was produced as described elsewhere [13]. The isotype-matching control MAb MOPC21 was acquired from MP Biochemicals. A family of 3E5 MAbs with identical V regions but different isotypes (IgG1, IgG2a, IgG2b, and IgG3) has been described elsewhere [14]. Actinium (²²⁵Ac) for construction of ²²⁵Ac/²¹³Bi generators was produced at the Institute for Transuranium Elements (Karlsruhe, Germany). Radiolabeling of MAbs with ²¹³Bi and ¹⁸⁸Re eluted from ¹⁸⁸Re/¹⁸⁸W generator (Oak Ridge National Laboratory, Oak Ridge, TN) was performed as described elsewhere [2]

In vitro killing of C. neoformanscells by²¹³Bi-18B7 and¹⁸⁸Re-18B7 MAbsFor RIT with ²¹³Bi, 5×10⁶ live cells were incubated with either labeled or unlabeled MAbs (0.2–4.0 μg/mL) for 0.5 h at 37°C, collected by centrifugation, incubated in PBS at 37°C for 3 h, treated with Tween 80 (0.5%), triturated 20 times, diluted, and plated for counting of colonies. For crossfire experiments, 5×10⁶ heat-killed (65°C for 1 h) C. neoformans cells were incubated with either labeled or unlabeled MAb (0.2–4.0 μg/mL) for 0.5 h at 37°C, collected by centrifugation, and mixed with 5×10⁶ live cells; the mixture was incubated for 3 h at 37°C, treated with Tween 80, diluted, and plated. The direct-hit and crossfire RIT experiments with ¹⁸⁸Re were conducted in essentially the same manner, except that C. neoformans cells were incubated for 48 h at 4°C, to prevent cell division before dilution and plating

Interaction of²¹³Bi-18B7 MAb with the C. neoformanscapsuleTreatment of H99 cells grown in capsule-inducing medium (hereafter, “induced H99 cells”) with ²¹³Bi-18B7 MAb was performed as described above, and the cell volume was measured as described elsewhere [12]

Binding of 18B7 MAb to H99 and 24067 cells¹⁸⁸Re-18B7 was added in different amounts to 24067, H99, or induced H99 cells (2×10⁶ cells/tube). After incubation for 1 h at 37°C, the samples were counted in a gamma counter, the cells were collected by centrifugation, and the pellets were counted. Scatchard analysis was used to compute the MAb affinity constant and the number of binding sites per cell, as described elsewhere [15]

Dependence of MAb binding to C. neoformanscapsule and C. neoformanskilling on MAb isotypeBinding of ²¹³Bi-3E5 MAbs to C. neoformans 24067 cells was performed as described above. The aliquots of cells treated with each of the labeled 3E5 MAbs were also tested for efficiency of C. neoformans killing

Radiation-triggered generation of ROSIndigo carmine bleaching was done as described elsewhere [10], with minor modifications. Approximately 2.5×10⁶C. neoformans cells were suspended in 10 mmol/L indigo carmine, and ²¹³Bi-18B7 (0.5 or 2.3 μCi/1 × 10⁵ cells), irrelevant MAb ²¹³Bi-MOPC21 (2.3 μCi/1 × 10⁵ cells), or unlabeled 18B7 (4.0 μg/mL) was then added. The absorbance of the reaction mixture at 610 nm was measured at 20-min intervals for 90 min. To provide a control, the cells were mixed with unlabeled 18B7 MAb, and the mixture was subjected to 254-nm UV light

Alternatively, we used an Amplex Red Hydrogen Peroxide/Peroxidase Assay Kit (Molecular Probes), which detects H₂O₂. MAb ²¹³Bi-18B7 (0.5–8.1 μCi/1×10⁵ cells) was added to 5 × 10⁶C. neoformans cells, followed immediately by the Amplex Red reagent and horseradish peroxidase in the reaction buffer. The absorbance of the reaction mixture at 540 nm was measured at 30, 50, and 85 min, and the amount of H₂O₂ produced was calculated

Quantification of the apoptosis-like death of C. neoformanscellsEither 5 mmol/L H₂O₂ (apoptosis-positive control) or 5 mmol/L ²¹³Bi-18B7 (0.6–2.4 μCi/1×10⁵ cells) was added to 2×10⁶C. neoformans cells, and the mixture was incubated for 3 h at 37°C, followed by an additional 3 h at room temperature. Untreated C. neoformans cells served as an apoptosis-negative control. The cells were subsequently stained with sulforhodaminyl-l-valylalanylaspartyl fluoromethyl ketone, which binds to all caspases, and fixed using a FLICA (fluorochrome-labeled inhibitors of caspases) kit (Immunochemistry Technologies), in accordance with the manufacturer’s instructions. Fluorescence was measured using a Becton Dickinson FACScan

Alternatively, the APO-BrdU TUNEL assay kit (Molecular Probes) was used in accordance with the manufacturer’s instructions. This method detects DNA fragments. Apoptosis-positive C. neoformans cells and apoptosis-negative control cells were prepared as described above. Also, the negative and positive controls (fixed human lymphoma cells) included in the kit were used for assay validation. The samples were stained with AlexaFluor 488 dye–labeled anti-BrdU antibody and propodium iodide and analyzed by flow cytometry

Macrophage experimentsMurine macrophage–like J774.16 cells were cultured and maintained as described elsewhere [15]. The viability of macrophages treated with 0–10 μCi of ²¹³Bi-18B7 for 3 h at 37°C was measured by trypan blue dye exclusion assay. To evaluate the interaction of RIT with macrophages, 1×10⁵ macrophages/well were plated on a 96-well plate and stimulated overnight with 500 U of interferon (IFN)–γ [16]. The medium was replaced with a suspension of 2×10⁴C. neoformans cells in fresh medium, giving a 10:1 effector-to-target cell ratio. C. neoformans cells were treated with different amounts of unlabeled 18B7 or with matching amounts of ²¹³Bi-18B7. After incubation for 3 h at 37°C, the supernatants from each well were collected, lysed, and plated for determination of colony counts. The effect of the medium on RIT was studied by performing RIT in PBS, Dulbecco’s modified Eagle medium (DMEM), and mouse serum

Cytokine analysisAll animal experiments were done in accordance with the guidelines of the Albert Einstein College of Medicine Institute for Animal Studies. Three groups of 10 AJ/Cr mice were infected intravenously with 1×10⁵C. neoformans H99 or 24067 cells. Twenty-four hours after infection, the mice either were left untreated or were treated intraperitoneally with 150 μCi of ²¹³Bi-18B7 (30 μg/mouse) or 30 μg of unlabeled 18B7. For the H99-infected mice, 5 from each group were killed at 48 h, and the fungal burden in their lungs and brains was analyzed. The remaining 5 mice in each group were killed on day 7 after infection, and their lungs and brains were homogenized in the presence of protease inhibitors (Complete Mini; Boehringer Mannheim) for cytokine analysis. The supernatants were tested for interleukin (IL)–2, IL-4, IL-10, tumor necrosis factor (TNF)–α, and IFN-γ by use of ELISA kits (BD PharMingen)

Flow-cytometric analysis of lung leukocytesThree groups of 5 AJ/Cr mice were infected intravenously with C. neoformans 24067 and either left untreated or treated (as described above) with ²¹³Bi-18B7 or unlabeled 18B7. On day 7 after infection, lungs were removed, homogenized, and digested enzymatically, and the cell suspension was collected by centrifugation as described elsewhere [8, 11]. For flow-cytometric analysis, the following MAbs were used: R-phycoerythrin–labeled anti-CD45, fluorescein isothiocyanate–labeled anti–mouse CD4, and CyChrome-labeled anti–mouse CD8 (BD PharMingen). The samples were analyzed using a Becton Dickinson FACScan. Controls consisted of isotype-matched irrelevant MAbs

Statistical analysisStudent’s t test for unpaired data was used to analyze differences in cytokine expression and the number of inflammatory cells between the treatment groups and to compare the results of the repeat experiments. Differences were considered to be statistically significant if P<.05. The Scatchard binding and ROS data were analyzed by linear and nonlinear regression (Prism; version 2; GraphPad)

Results

Killing of C. neoformansby²¹³Bi-labeled MAb through both direct-hit and crossfire effectsTo elucidate the contribution of direct-hit and crossfire effects to RIT for C. neoformans infection, we compared the fungicidal activity of MAb conjugated to ²¹³Bi and ¹⁸⁸Re, isotopes with different emission ranges in tissue (50–80 μm for ²¹³Bi vs. 10 mm for ¹⁸⁸Re). In cancer RIT, it is assumed that ²¹³Bi kills through direct-hit effects and that ¹⁸⁸Re kills through crossfire effects [3, 17, 18]. In principle, every cell with bound radiolabeled MAb molecules can be killed by a direct hit and simultaneously serve as a source of crossfire radiation. By measuring the killing of cells during RIT and in crossfire experiments, we can calculate the contribution of direct-hit effects toward cell killing by subtracting the percentage of cells killed by crossfire effects from the percentage of cells killed by RIT

To observe crossfire, we had to ensure that the cells that served as the sources of crossfire radiation could not themselves be killed by direct hit. Consequently, we used heat-killed C. neoformans cells. Experiments with ²¹³Bi-18B7 showed that, although most kills were due to direct hits, the crossfire effect also contributed to the fungicidal effect of RIT (figure 1B). No killing of C. neoformans cells by unlabeled 18B7 was observed. We have previously shown that radiolabeled irrelevant MAbs were unable to kill C. neoformans cells [2]. For ¹⁸⁸Re-18B7, the crossfire effect was responsible for most of the C. neoformans killing (figure 1C)

Reduction of capsule size by treatment of C. neoformanscapsule with capsule-binding²¹³Bi-18B7 MAbIncubation of C. neoformans with ²¹³Bi-18B7 resulted in a significant reduction (P<.001) in capsule size (figure 2). This effect was not observed with unlabeled 18B7

Effect of strain and MAb isotype on capsule binding and cell killing by RITWe measured pronounced differences in the abilities of 24067, H99, and induced H99 cells to bind to 18B7 MAb (figure 3A) and found lower affinity constants for 18B7 MAb binding to H99 and induced H99 cells than for binding to 24067 cells (table 1). Strains 24067 and H99 differed significantly in their susceptibility to RIT (figure 3B). The ²¹³Bi-3E5 MAbs bound differently to 24067 cells (figure 3C), with 3E5 IgG1 and 3E5 IgG2b having the lowest and highest affinity constants, respectively. The killing of C. neoformans cells by ²¹³Bi-labeled MAbs was proportional to their affinity constants (figure 3D) and was described by the 1-phase exponential decay equation Y=398e^-0.38x+18.71 (R²=0.8)

Induction of the formation of ROS by radiationExposing the mixture of C. neoformans cells and unlabeled 18B7 to UV light initiated the formation of singlet molecular oxygen (¹O₂*), which is required for MAb-catalyzed production of ROS [10] (figure 4A). Radiation from radiolabeled MAb in solution, which causes the hydrolysis of water, also triggered the antibody-catalyzed generation of ROS. For ²¹³Bi-18B7, ROS formation was dose dependent, with ROS formation observed at 2.3 μCi/1 × 10⁵ cells and no ROS formation for 0.5 μCi/1 × 10⁵ cells (the slope was not significantly different from zero; P=.37) (figure 4B). No ROS formation was observed for the control ²¹³Bi-MOPC21 (P=.20), suggesting a need for antigen-antibody interaction in this process (figure 4B). The indigo carmine bleaching technique was validated by use of Amplex Red reagent, which in the presence of peroxidase reacts with H₂O₂ to produce red-fluorescent resorufin [19]. The amount of H₂O₂ produced was clearly dependent on the level of radioactivity in the sample (figure 4C)

Induction of apoptosis-like death in C. neoformanscells by RITEnvironmental stresses, antibiotics, and UV light can induce apoptosis-like death in fungal cells [20–22]. RIT caused a significant percentage of C. neoformans cells to undergo apoptosis-like death, as determined by FLICA assay [23], and the effect was dose dependent, with a mean ± SD of 8%±3% and 18%±5% of cells demonstrating apoptosis-like changes after treatment with 0.6 μCi and 1.2 μCi of ²¹³Bi-18B7, respectively (P=.03). Treatment of C. neoformans cells with 5 mmol/L H₂O₂ resulted in 20% of cells demonstrating apoptosis-like changes, which matches published results for Candida albicans [20]. At 2.4 μCi/1×10⁵ cells, the percentage of cells demonstrating apoptosis-like changes remained the same (mean ± SD, 18%±6%; P=.9), possibly because some treated cells became necrotic. When cells are exposed to high concentrations of apoptosis-causing agents (such acetic acid and H₂O₂), necrotic cells can outnumber apoptotic cells, which results in reduced staining [20]. FLICA data was validated by the TUNEL technique, using the APO-BrdU TUNEL assay kit [24]. The mean ± SD percentages of apoptotic-like cells as determined by the TUNEL technique were 10%±5%, 20%±8%, and 22%±10% for 0.6, 1.2, and 2.4 μCi of ²¹³Bi-18B7/1×10⁵ cells, respectively, with the difference between the groups treated with 0.6 and 1.2 μCi being significant (P=.04)

Cooperation of macrophages with²¹³Bi-labeled MAb in killing of C. neoformanscellsInitially, we determined whether radiolabeled MAb affected the viability of macrophages. No effect on macrophage viability was seen with doses of up to 8 μCi of ²¹³Bi-18B7 (data not shown). However, the presence of ²¹³Bi-18B7 significantly enhanced the killing of C. neoformans by J774.16 cells, compared with that seen with unlabeled MAb (figure 5A), and the process was less efficient in DMEM and mouse serum than it was in PBS (figure 5B)

Reduction of expression of cytokines in vivo by RITThere was a 2-fold reduction of fungal burden in the lungs and brains of RIT-treated mice infected with strain H99 (mean ± SD, 1.5±0.6 and 2.5±1.1×10⁵ cfu/g of tissue, respectively), compared with that seen in control groups (mean ± SD, 3.4±1.0 and 5±2.0×10⁵ cfu/g of tissue, respectively) (P=.03). Hence, H99 strain was susceptible to killing with radiolabeled MAb in vivo, but the efficacy was lower than reported in our previous experiments with the 24067 strain [2]. RIT was associated with changes in cytokine expression in the lungs for both the H99 strain (figure 6A and 6B) and the 24067 strain (figure 6C). For the H99 strain, there was a significant reduction (P<.05) in the expression of TNF-α, INF-γ, and IL-10 (brains and lungs), IL-4 (lungs), and IL-2 (brains) in the RIT-treated mice, compared with that in the untreated control mice. There was a trend toward increased IL-2 expression in the lungs of the RIT-treated mice (P=.06); also, unlabeled 18B7 MAb caused some increase in IL-4 expression in the brains of the mice. For the 24067 strain, a significant reduction (P<.02) in the expression of IL-2, IL-4, and IL-10 in both the lungs and brains of the RIT-treated mice was detected, compared with that in the untreated control mice

Association between RIT and fewer inflammatory cells There were significantly fewer CD45⁺ cells in the lungs of mice treated with unlabeled MAb or ²¹³Bi-18B7 than in the untreated control mice (table 2). No difference in CD45⁺ cell numbers was observed between the unlabeled MAb and ²¹³Bi-18B7 groups. Significantly fewer CD4⁺ cells were found in the unlabeled MAb and ²¹³Bi-18B7 groups than in the untreated control mice, whereas the decrease in CD8⁺ cells in the ²¹³Bi-18B7–treated mice did not achieve statistical significance. However, in the unlabeled 18B7–treated mice, there was a significant decrease in the number of CD8⁺ cells

Discussion

There is a wealth of preclinical and clinical information about interactions between radiolabeled antibodies and tumor cells and about RIT-associated toxicity [3, 6, 25–28]. For example, a reduced therapeutic effect against breast cancer and lymphoma has been observed when Fc receptor activation was absent [29]. However, no comparable information is available for RIT for infectious diseases

The C. neoformans MAb system permits exploration of fundamental problems in radiobiology that have not been addressed previously. Because MAb binding to C. neoformans is not complicated by internalization of immunoglobulin, we were able to separate direct-hit and crossfire effects by using heat-killed cells with bound ²¹³Bi-18B7 as a source of crossfire radiation. For ²¹³Bi-18B7, the mechanism of killing involved both direct-hit and crossfire effects, whereas, for ¹⁸⁸Re, killing was caused almost exclusively by crossfire effects, a finding that is consistent with the radiological properties of these isotopes. If the direct-hit effect is the major mechanism for fungal cytotoxicity with ²¹³Bi-18B7, then one would expect efficacy to be a function of affinity, because the isotope needs to be delivered in close proximity to the target cell. Consistent with this prediction, cell-killing experiments with ²¹³Bi-3E5 MAbs revealed that killing was proportional to the affinity constant

Another mechanism that could potentially contribute to the killing of C. neoformans by RIT is the formation of ROS, an effect that may be catalyzed by antibodies [10]. However, ROS are unlikely to contribute significantly to fungal-cell killing, because no significant oxidant formation was observed for the dose of 0.5 μCi/1×10⁵ cells, which was fungicidal

The polysaccharide capsule can mediate many deleterious effects on host immune responses [30]. Previously, we showed that ¹⁸⁸Re-18B7 binding to C. neoformans reduced the capsule volume of H99 cells [31]. Here, we demonstrate that ²¹³Bi-8B7 has a similar effect, thus extending this observation to α-radiation. Reduction in capsule size may decrease the virulence of individual cells and, thus, contribute to the therapeutic effect of RIT. The susceptibility of different C. neoformans strains to RIT was clearly dependent on the ability of a given strain to bind to the radiolabeled MAb (figure 3 and table 1), with 24067 (for which the 18B7 affinity constant was the highest) being more susceptible to RIT than H99

Environmental stresses, antibiotics, and UV light can induce apoptosis-like death in fungal cells [20–22]. Here, we show that radiation can induce dose-dependent apoptosis-like changes in C. neoformans. Consequently, radiation-induced apoptosis-like events may contribute to the efficacy of RIT for infectious diseases in a manner similar to the RIT effects on tumor cells [6]

Macrophages remained viable in the presence of radiolabeled MAb, but the efficacy of fungal killing by J774.16 cells was significantly enhanced by ²¹³Bi-18B7, compared with that seen with unlabeled MAb. This cooperation was impressive in view of the fact that the efficacy of C. neoformans killing by ²¹³Bi-18B7 alone in DMEM or in mouse serum was reduced, compared with the efficacy of killing in PBS. Serum and DMEM contain sugars, proteins, and vitamins that may attenuate the fungicidal effects of RIT. It is also possible that the binding of radiolabeled MAb to C. neoformans cells might be reduced in DMEM and serum

Given that MAb-mediated protection in the setting of C. neoformans infection may be in part due to changes in cytokine expression [8, 11], we investigated whether RIT affected IL-2, IL-4, IL-10, TNF-α, and IFN-γ expression in RIT-treated infected mice. These cytokines are representative of the Th1- and Th2-related cytokines that are important in the effective immune response to C. neoformans infection [8, 32, 33]. The reduction in the organ fungal load observed in the present experiments in strain H99–infected mice was lower than that previously observed for strain 24067–infected mice [2], which is likely to be the direct consequence of decreased binding of radiolabeled 18B7 to H99 cells and decreased killing by radiation. Ben-Yosef et al. [34] observed a reduction in the number of colony-forming units in the brains of mice infected systemically with C. neoformans and subjected to low doses of external γ-radiation. One possible explanation of the reduction in cytokine expression might be decreased inflammation in the organs as a result of decreased fungal burden. However, it is also plausible that some of the inflammatory cells were killed by crossfire radiation from the radiolabeled MAb. The reduction in cytokine expression was the same for both the H99 strain and the 24067 strain, and the percentage of change did not correlate with the change in microbial load. We note that there was no significant difference in the total number of CD45⁺ cells in the lungs of mice treated with labeled and nonlabeled 18B7, which would argue against wholesale killing of inflammatory cells at sites of infection. However, it is conceivable that local radiation could have affected the number or location of inflammatory cells, thus resulting in an alteration in cytokine expression. In this regard, external ionizing radiation can lower CD4⁺ and CD8⁺ cell counts in radiation workers and in patients with cancer undergoing radiation therapy [35, 36]. Given the hypothesis that the death of C. neoformans–infected mice can be caused by host inflammatory damage [32], the lower percentage of CD4⁺ cells in RIT-treated mice raises the tantalizing possibility of benefit from reduced inflammatory damage

In summary, we identified 4 potential mechanisms by which RIT may affect the outcome of C. neoformans infection in treated mice: direct fungicidal effects that may follow promotion of apoptosis-like events, reduction in capsule size, enhancement of C. neoformans killing by macrophages, and subtle changes in the inflammatory response. In contrast, we could not implicate ROS in the fungicidal effects. Hence, the beneficial effects of RIT are probably mediated through a variety of mechanisms. The finding that affinity correlated with efficacy strongly suggests that higher-affinity MAbs would be superior reagents for use in RIT