SYTO-13, a Viability Marker as a New Tool to Monitor In Vitro Pharmacodynamic Parameters of Anti-Pneumocystis Drugs

While Pneumocystis pneumonia (PcP) still impacts the AIDS patients, it has a growing importance in immunosuppressed HIV-negative patients. To determine the anti-Pneumocystis therapeutic efficacy of new compounds, animal and in vitro models have been developed. Indeed, well-designed mouse or rat experimental models of pneumocystosis can be used to describe the in vivo anti-Pneumocystis activity of new drugs. In vitro models, which enable the screening of a large panel of new molecules, have been developed using axenic cultures or co-culture with feeder cells; but no universally accepted standard method is currently available to evaluate anti-Pneumocystis molecules in vitro. Thus, we chose to explore the use of the SYTO-13 dye, as a new indicator of Pneumocystis viability. In the present work, we established the experimental conditions to define the in vitro pharmacodynamic parameters (EC50, Emax) of marketed compounds (trimethoprim/sulfamethoxazole, pentamidine, atovaquone) in order to specifically measure the intrinsic activity of these anti-P. carinii molecules using the SYTO-13 dye for the first time. Co-labelling the fungal organisms with anti-P. carinii specific antibodies enabled the measurement of viability of Pneumocystis organisms while excluding host debris from the analysis. Moreover, contrary to microscopic observation, large numbers of fungal cells can be analyzed by flow cytometry, thus increasing statistical significance and avoiding misreading during fastidious quantitation of stained organisms. In conclusion, the SYTO-13 dye allowed us to show a reproducible dose/effect relationship for the tested anti-Pneumocystis drugs.


Introduction
Pneumocystis pneumonia (PcP) still has a high impact on AIDS patients from developed or developing world areas because of newly diagnosed HIV infections, non adherence or limited access to either highly active antiretroviral therapy (HAART) or PcP chemoprophylaxis [1]or high contribution to AIDS mortality [2]. In HIV-negative patients submitted to immunosuppressive treatments for malignancies, auto-immune diseases or organ transplantation, PcP is the cause of pneumonia in 10%-40% of these patients, with high mortality rates [3,4]. In addition, Pneumocystis organisms were detected relatively frequently in neonates or small children [5,6], pregnant women [7] and patients with chronic underlying diseases [8][9][10]. Indeed, in the latter population, Pneumocystis organisms have been considered as co-morbidity factors worsening the prognosis [3,5,11,12]. Particularly, in patients with chronic obstructive pulmonary disease (COPD), a colonization of the lungs by P. jirovecii is associated with a more progressive and severe disease [12,13,14]. In clinical practice, the therapeutic choices are limited due to the lack of therapeutic alternatives to the well-known trimethoprim-sulfamethoxazole (TMP/SMX) drug combination and warrant the search for more effective and less toxic agents. Moreover, some recent reports have suggested the emergence of P. jirovecii resistance to sulpha drugs [15,16].
The therapeutic efficacy of new compounds against Pneumocystis micromycetes, is usually determined using well-defined mouse or rat experimental models of pneumocystosis [17][18][19][20][21] or/and in vitro models.
Since no continuous culture system supporting the replication of the organism is available, in vitro models, which enable the screening of a large panel of new molecules, have been developed using axenic cultures or co-cultures with feeder cells [22][23][24]. However, no universally accepted standard method is currently available to evaluate anti-Pneumocystis molecules in vitro. The main reasons for this situation are (i) the microscopic observation of Pneumocystis is time consuming and requires a great expertise [25,26] and (ii) the assessment of Pneumocystis viability with high confidence and reproducibility is lacking [27]. Many methods have been described such as the incorporation of classical vital stains or fluorescent indicator compounds [26,28,29], the ATP bioluminescent assay [27,30], the uptake of radiolabelled methionine, uracil or thymidine [31], and the RT-PCR [32,33]. However, none of these overcomes the host cell debris interferences and none has permitted to monitor the concentration-effect relationships (i.e. pharmacodynamics parameters) of anti-Pneumocystis compounds. As an alternative method, we evaluated the use of SYTO-13, a member of the class of SYTO dyes, as a new indicator of Pneumocystis viability. The SYTO dyes, among which is the most popular SYTO-13 [34], are cell-permeant nucleic acid stains that permit myriad of applications such as the discrimination between live/dead eukaryotic or prokaryotic cells [35][36][37], the detection of apoptosis [38,39], or germinated bacterial endospores [40].
The aim of the present work was to establish the in vitro experimental conditions allowing the definition of the in vitro pharmacodynamic parameters of three marketed compounds (TMP/SMX, pentamidine, atovaquone). For the first time, the SYTO-13 dye enabled to specifically characterize the intrinsic anti-Pneumocystis activities of these molecules and to show a reproducible dose/effect relationship, while being more reproducible and avoiding laborious microscopic observations.

Results
Labelling of P. carinii organisms with the SYTO-13 live-cell nucleic acid stain P. carinii was stained using a specific anti-Pneumocystis polyclonal antibody and a goat anti-rat IgG conjugated to Alexa-647 (Fig 1). Pneumocystis trophic and cystic forms are both labelled in red. The host lung debris were not labelled by the polyclonal antibody. SYTO-13, a New Viability Marker for Pneumocystis SYTO-13 is a sensitive DNA stain broadly used for viability studies. SYTO-13 dye easily penetrates most cell types and undergoes dramatic fluorescence enhancement upon binding to nucleic acids. Thus, when P. carinii was incubated with SYTO-13, the nuclei from living cells were stained in green (Fig 1). Cells that were stained in red but did not display any nuclear green fluorescence were considered as non-viable P. carinii organisms (Fig 1C, 1G and 1K).

P. carinii viability assessment
To specifically measure the viability of Pneumocystis organisms, a first step consisted in selecting the labelled P. carinii population using flow cytometry (Fig 2). Thus, the gated population R1 (Fig 2A) allowed the isolation of P. carinii organisms from the host pulmonary cell debris. Then, the SYTO-13 fluorescence signals were analyzed within R1. Fig 2B represents the viable population (R2 gate) of untreated P. carinii after 4 days of culture in DMEM with 10% FCS. The percentage of viability reached 72.00 ± 7.02% (n = 8) for untreated cultured Pneumocystis organisms. When P. carinii cells were incubated with increasing concentrations of anti-Pneumocystis drugs, the number of viable cells (R2) decreased gradually until hardly no viable P. carinii organism was detected: as an example, 56.70 ± 1.56% (n = 3) and 8.60 ± 0.04% (n = 3) of viable P. carinii organisms for 0.15 μM and 90 μM of pentamidine, respectively ( Fig 2C and 2D).

Pharmacodynamic parameters
The inhibition of P. carinii viability in the presence of anti-Pneumocystis drugs was compared to drug-free control using the SYTO-13 assay (Fig 3). The concentration-response curves were plotted after 4 days of incubation of P. carinii organisms with pentamidine, atovaquone or TMP/SMX. The reduction in the number of viable microorganisms was gradual and dependent of drug concentration. Data were integrated using the Hill equation (sigmoid Emax model) to calculate the pharmacodynamic parameters of each tested drug (Table 1) (Table 1). In comparison, pentamidine showed a maximum killing rate of P. carinii with an EC50 value of 0.45 μM and atovaquone, with an EC50 value of 1.89 μM. In terms of efficacy, pentamidine and atovaquone reached at least 90% of inhibition at a concentration of 145 μM. Higher concentrations of TMP/SMX (1/6 mM) were needed to achieve 90% of inhibition. Finally, our data showed that the potency of atovaquone and TMP/SMX were very sensitive to variations of drug concentration, as attested by the important steepness of the curves (4.00 and 2.46, respectively, Table 1).

Discussion
The availability of a robust and consistent method to assess the intrinsic activity of drugs is a key step in the development of new anti-Pneumocystis treatments. A pharmacodynamic model was previously developed to that purpose [20,41]. Thus, the in vitro drug activities were determined using the broth microdilution technique, comparing the total number of microorganisms in treated and drug-free cultures as monitored after Giemsa staining. The in vitro maximum effect (Emax, efficacy), the drug concentrations for which 50% of Emax is obtained (EC50, potency), and the slope of the dose-response curves were then calculated by the Hill equation (sigmoid Emax model). However, two drawbacks arise from this methodology: (i) the microscopic quantitation of Pneumocystis organisms is time-consuming and requires a great expertise; (ii) the Giemsa staining does not enable to distinguish viable from non-viable Pneumocystis organisms although assessing the viability remains a key feature to evaluate antimicrobial activity accurately.
Most viability assays using the light microscope are not really applicable to Pneumocystis. Indeed, the interpretation of results acquired with most staining methods is often inconclusive. Only a double-staining fluorogenic method [28,29] was considered to be reliable to evaluate the viability of Pneumocystis fungal cells. This method employs a carboxyfluorescein diacetate labelling, which positively stains live parasites in fluorescent green, combined with propidium iodide, that intercalates into double-stranded nucleic acids, and stains only dead parasites in fluorescent red. Other authors used propidium iodide alone to assess the viability of P. carinii by flow cytometry analysis [26]. Viability was also evaluated by monitoring metabolic activities such as the de novo synthesis of folates [42], the incorporation of para-aminobenzoic acid (pABA) [43], or the uptake of methionine, uracil, thymidine [31] or dimethyaminostyrylmethylpyrimidinium-iodide (DASPMI) [44]. Moreover, some authors used detection of Heat Shock Protein 70 transcripts by RT-PCR to monitor the viability of P. jirovecii in bronchoalveolar lavage (BAL) fluids from patients developing PcP [32,33] or in air samples collected from the environment of PcP patients [33]. However, the concentration-effect relationship (i.e. pharmacodynamic parameters) of anti-Pneumocystis compounds has not been characterized by these methodologies. Moreover, the viability assays based on microscopic observations are limited because of the low numbers of analyzed cells and potential interferences due to the host cell debris.
In the present work, SYTO-13, a member of the class of SYTO dyes, is proposed as a new tool to monitor Pneumocystis viability. The SYTO live-cell nucleic acid stains are sensitive DNA stains that are broadly used for live-cell studies [35][36][37]. The SYTO dyes readily penetrate most cell types and undergo dramatic fluorescence enhancement upon binding to nucleic acids as we have shown it in our experiments (Figs 1 and 2). The SYTO-13 dye has proven valuable in various fields of research applications such as (i) live/dead discrimination in eukaryotic or prokaryotic cells [35][36][37], (ii) detection of apoptosis [38,39] and (iii) quantification of bacterial endospore germination [40].
In our hands, SYTO-13 also revealed to be an excellent dye to assess the effect of anti-P. carinii drugs. Indeed, the results concerning the response to standard anti-Pneumocystis drugs obtained in our pharmacodynamic in vitro models using SYTO-13 were consistent with previous in vitro efficacy studies [20,41]. However, the EC50 values calculated after Giemsa staining (using the same procedure and Pneumocystis strain detailed in references [20,41]) are 2-to 6-fold higher than those obtained with SYTO-13 highlighting the difficulty to distinguish viable P. carinii organisms by microscopic observations. In other words, there is a tendency to over-estimate the amount of dead P. carinii after Giemsa staining.
SYTO-13 allowed the accurate follow-up of the shift of P. carinii organisms from the viable to the dead compartment upon increase of drug concentrations (Fig 2B, 2C and 2D). Thus, it became feasible to calculate the percentage of viability inhibition in relation to untreated P. carinii control. By using anti-P. carinii specific antibodies, we developed a specific and sensitive tool to assess the viability of P. carinii organisms in the presence of marketed compounds, while selecting fungal organisms against host cell debris. Moreover, a large number of fungal cells was analyzed by flow cytometry leading to consistent and objective results, while avoiding misreading or false interpretation that often occurs when observing Giemsa-stained and rather small (2-8 μm in diameter) P. carinii organisms under the microscope.
In conclusion, the SYTO-13 staining assay combined with a flow cytometry analysis appears as a simple and reliable approach for the measurement of Pneumocystis viability. We have also demonstrated that SYTO-13 is compatible with the application of the pharmacodynamic Emax model, enabling to show a reproducible dose/effect relationship for the tested anti-Pneumocystis marketed drugs. Finally, it could be interesting to investigate the use of SYTO-13 to measure the viability of P. jirovecii collected from patients with PcP and subsequently cultured in the presence of drugs for a short period of time in order, for example, to explore the potency of drug treatment.

Source of P. carinii organisms
Athymic Pneumocystis-free Lou nu/nu rats (Pasteur Institute Lille, France) were used as source of Pneumocystis carinii organisms for all experiments [45]. Ten-week-old female nude rats were administered dexamethasone (Merck Sharp & Dohme Chibret, Paris, France) for two weeks in the drinking water (1 mg/L). Then, rats were inoculated with 10 7 of cryopreserved parasites using a non-surgical endotracheal method [18]. Dexamethasone treatment was maintained until the end of the experiment. Six to 8 weeks post-inoculation (p.i.), rats were highly infected by P. carinii, without secondary fungal or bacterial infection. Animals were housed in HEPA-filtered air isolators (Flufrance, Wissous, France) and were allowed sterile irradiated food (Scientific Animal Food & Engineering, SAFE, Augy, France) and sterile water ad libitum.
Extraction, purification and quantitation of P. carinii organisms Six to 8 weeks following inoculation, rats were euthanatized and parasite extraction was performed as previously described [20]. Briefly, parasites were extracted in DMEM (BioWhittaker, France) by agitation of lung pieces with a magnetic stirrer. The resulting homogenate was poured successively through gauze, 250 and 63 μm stainless steels filters. After centrifugation, the pellet was suspended in an haemolytic buffered solution. P. carinii organisms were collected by centrifugation and then purified on a polysucrose gradient (Histopaque-1077, Sigma-Aldrich, France). Blood and Sabouraud dextrose agar (Difco, France) media were inoculated with purified parasites to check for the absence of contaminating pathogens. P. carinii was quantitated on air dried smears stained with RAL-555 (Réactifs RAL, Martillac, France), a rapid panoptic methanol-Giemsa-like staining, which stains trophic forms, sporocytes and cysts of P. carinii (20). P. carinii was then cryopreserved by placing parasites in FCS with 10% DMSO at -80°C in a Nalgene 1°C cryo-freezing container (cooling rate: about 1°C/min) for 4 hours [46]. The parasite samples were then stored in liquid nitrogen. Cryopreserved P. carinii were used for in vitro anti-microbial studies.
Axenic in vitro culture of P. carinii In order to determine in vitro drug susceptibility of P. carinii, axenic cultures of the organism were performed as follows [20]. All the experiments were carried out in 24-well plates with a final volume of 2 mL of DMEM supplemented with 10% FCS containing a final inoculum of 10 6 organisms per mL. Plates with organisms were incubated for 4 days in an atmosphere of 5% CO 2 at 37°C. Then, the total volume of each well was removed, centrifuged for 10 min at 2,900 × g and the pellet was suspended with 100 μL of phosphate buffer solution (PBS) Dulbecco (Sigma Chemical Co.).

In vitro susceptibility studies
In vitro susceptibility studies were performed using the twofold broth microdilution technique. Final drug concentrations ranged from 150 μM to 1.5 nM for pentamidine and atovaquone.
TMP/SMX combination was tested from 1.7/10 mM to 3.5/20 μM. Plates were incubated during 4 days in an atmosphere of 5% CO 2 at 37°C. Viability of Pneumocystis organisms incubated in such culture conditions, but without any drugs, was assessed using the SYTO-13 assay and was set to 100% of viability. Such a control is named the drug-free control. All susceptibility assays were set up in triplicate.

Pneumocystis labelling
An immunofluorescence assay (IFA), using a specific anti-Pneumocystis polyclonal antibody, was performed directly in each 100μL-parasite culture suspension, prior analysis by flow cytometry. We used a specific anti-Pneumocystis polyclonal antibody made in our laboratory as follows: P. carinii endotracheally-inoculated Sprague Dawley rats (Harlan, France) were immunosuppressed for 10 weeks with dexamethasone (2 mg/L in the drinking water). The infected animals produced high Pneumocystis loads. Then, animals received half of the dexamethasone dose for one week before stopping the immunosuppression treatment. After 8 to 10 weeks of recovery, the antibody titer was determined in the collected serum. The P. carinii trophic and cystic forms were labelled with high intensity by the polyclonal antibody as checked by IFA. A goat anti-rat IgG (H+L) conjugated to Alexa-647 (Invitrogen Life Sciences, Cergy Pontoise, France) was used as secondary antibody. The optimal dilutions used for the polyclonal antibody and the secondary antibody were 1:100 and 1:50, respectively. The organisms were washed once with PBS Dulbecco (10 min at 2,900 × g) between each antibody incubation steps. Antibodies were incubated 30 min at 37°C.
Pneumocystis viability using SYTO-13 live-cell nucleic acid stain After Pneumocystis labelling with the specific polyclonal antibody, each 100μL-parasite culture suspension was incubated at room temperature with 1 mL of SYTO-13 (Molecular Probes Europe BV, Leiden, The Netherlands). Different concentrations of SYTO-13 (1 to 20 μM) and different incubation times (10 to 60 min) have been tested. The optimal concentration to label Pneumocystis organisms were found to be 2.5 μM of SYTO-13 in PBS Dulbecco with an incubation time of 20 min.

Flow cytometry analysis
Immunostained P. carinii organisms, that were incubated with SYTO-13, were analyzed using the FACScalibur 2 flow cytometer (Becton Dickinson) driven by the BD CellQuest software (version 0.3.df6b, Becton Dickinson). The cytometer is equipped with an air-cooled blue laser providing 15 mW at 488 nm and the standard filter setup. All parameters were collected as logarithmic signals. Green fluorescence (SYTO-13 staining) was collected in the FL1 channel (530 ± 30 nm) whereas red fluorescence (Alexa-647 goat anti-rat IgG; P. carinii staining) was collected in the FL4 channel (661 ± 16 nm). The SYTO-13-stained P. carinii organisms were numbered within the gate selecting for the anti-P. carinii polyclonal antibody positive staining. The viable P. carinii are gated (R2, Fig 2) inside the cell population with high green fluorescence and the percentage of viability is then calculated in comparison with all events. The in vitro activity of tested compounds against P. carinii was expressed as a percentage of viability inhibition defined as the total number of SYTO-13 positive parasites for a given drug concentration in comparison with the number of untreated parasites.

Determination of pharmacodynamic parameters
Once the percentages of inhibition of P. carinii viability were calculated, the Hill equation was applied to establish the relationship existing between the concentration and the inhibitory effect of a given drug as follows [47]: E R = (E R,max x C S ) / [(EC 50 ) S + C S ]. E R is the effect of each drug concentration on the percentage of inhibition estimated from experimental results (C); S is a parameter reflecting the steepness of the concentration-effect relationship curve; EC 50 is the concentration of the compound at which 50% of the maximum effect (E R,max ) is reached.
The parameters of this pharmacodynamic model were calculated by nonlinear least-square regression techniques using a commercial software (WinNonlin).