Ribosome inactivating proteins (RIPs) irreversibly damage ribosomes [1
], leading to inhibition of protein synthesis and cell death [2
]. Ricin and abrin—two well-known plant-derived toxins—are RIPs with similar structures, comprised of two glycoprotein chains (A and B) of equal sizes (~30 kDa) joined by a disulfide bond. The B-chain binds to galactose residues present on various cell surface glycoproteins and glycolipids, triggering endocytosis of the toxin. The A-chain exhibits a ribonucleic acid (RNA) N-glycosidase activity which depurinates a specific adenine residue located near the 3′ terminus of the 28S ribosomal RNA. This site-specific depurination event prevents binding of elongation factor 2 to the ribosome, thereby causing translational arrest [3
]. Ricin and abrin are isolated from the seeds of castor bean plant Ricinus communis
and Abrus precatorius
, respectively. The ease and low cost of isolation make these proteins potent biothreat agents.
Currently, there are two main approaches for assessing ricin activity in noncell-based assays: measuring protein synthesis inhibition [5
] and monitoring ribosomal 28S RNA depurination [9
]. Although highly sensitive, these approaches focus on measuring the catalytic activity of ricin subunit A. Yet, to accurately assess the full activity of ricin toxicity, a cell-based assay should be employed. Current reported cell based assays for ricin toxicity measure cell survival (usually within 48 to 72 h postexposure) [16
] or measure intracellular marker protein levels within 3–24 h postintoxication [6
Cell death induction by apoptosis, necrosis, and autophagy is characterized by a distinct set of temporal morphological, biochemical, and gene expression features [21
]. These specific morphological features, in particular changes in cell volume, accompany cell death processes. Thus, they are often used to differentiate between different cell death pathways—for example, loss of cell volume or shrinkage are morphological characteristics of apoptosis, while initial cell swelling is a characteristic of necrosis [22
]. Several studies have shown that the cell death resulting from ricin and abrin intoxication is apoptotic in nature, induced by activating caspase-3
. This has been demonstrated in several cancer cell lines as well as mammalian lymphoid tissues [23
There are several commercial in vitro cell-based assays for assessing cell viability by the loss of membrane integrity. Such assays include nuclear staining of nonviable cells by trypan blue or propidium iodide dying [25
] or measuring lactate dehydrogenase in the surrounding medium [27
]. These assays are more pronounced and accurate at the late stages of cell death, but are destructive to the sample, require specific reagents and several hours for completion and are retrospective in their analysis. Time-lapse experiments for living cells using fluorescent microscopy are challenging to implement routinely and require substantial resources [28
]. While signal specificity obtained from fluorescent microscopy is considered a major advantage, it suffers from some intrinsic limitations: requirement of exogenous labels, invasive procedures, appropriate reagents, and time-consuming image focusing process prior to image acquisition and analysis.
Digital holographic microscopy (DHM) is a time-lapse, label-free, and noninvasive imaging technique that can provide both quantitative and qualitative phase information on morphological features such as cellular area, shape, thickness, volume and confluence [29
]. This method does not require any media changes or addition of dyes, simplifying high throughput screening. Thus, this technology is suitable for sensitive measurements of various cellular events such as live cell imaging, cell migration, proliferation, differentiation, and death [31
]. Identifying and measuring specific morphological features by DHM during early stages of exposure to toxic agents such as ricin and abrin, could allow early cell death detection, thereby determining toxin activity. As a label-free technique, DHM does not require pipetting steps which can affect the cell and even induce strong bias as some detached dead cell can be washed out. We therefore set out to develop an activity assay for ricin and abrin based on DHM evaluation of cells following intoxication.
In this study, we present the feasibility of a cell-based assay for ricin and abrin intoxication using DHM. In contrast to current methodologies, DHM is noninvasive, rapid and able to monitor changes in cell morphology dynamically, allowing early detection of low dose intoxication. The method described herein may be applied for early and sensitive detection of these as well as other toxins and substances.
Digital holographic microscopy (DHM) is a time-lapse, label-free, and noninvasive imaging technique that can provide both qualitative and quantitative phase information on morphological features of cells. Detection by DHM of morphological features characteristic of early stages of cell death [37
] can provide a rapid and sensitive method for assessing cell cytotoxicity upon interaction with drugs, toxins and interfering compounds.
In this study, we employed DHM to evaluate the morphological changes of cell lines during ricin and abrin intoxication. We demonstrated that the effect of ricin and abrin intoxication on HeLa and Vero cell lines is characterized by a decrease in cell confluence which was also observed by monitoring less cell divisions and a reduction of total cell counts compared to untreated cells. In addition, a decrease in morphological parameters such as cell area, perimeter, irregularity, and ECV, and an increase in optical thickness and roughness were also observed. These effects are directly attributed to ricin or abrin intoxication as they were inhibited by the addition of neutralizing antibodies against ricin or abrin respectively.
The morphological changes analyzed by the DHM are related to the total cell population. To exclude the contribution of dead cells from the analysis we used the cell tracking application of the DHM. This allowed us to examine the effect of ricin on live cells exclusively. Detection of morphological changes in presumed live cells upon early stages of intoxication reflects a rapid and sensitive method for assessing toxin activity. It was found that HeLa cells exhibited earlier significant morphological changes following intoxication and a significant reduction in cell viability, when compared to Vero cells.
Shortening the preparation time of the assay by exposing preadherent HeLa cells to toxins increased assay sensitivity, with an earlier onset of detectable morphological changes compared to fully adhered cells. A further reduction of toxin concentrations was performed in a dose–response assay (0.1–100 ng/mL), showing clear detection of morphological changes within 4 to 6 h for 10 ng/mL toxins. Significant changes in confluence and phase-shift parameters were also observed in 0.3 ng/mL intoxication within 15 h, defining the limit of detection of the assay. The effect of cell adhesion on drug- induced apoptosis was well described in several studies [42
]. Damiano et al. (1999) demonstrated the critical role of integrins and extracellular matrix interaction in cell survival and showed that cells in direct contact with the integrin fibronectin are less sensitive to acute doxorubicin treatment.
The addition of anti-ricin polyclonal antibodies induced a complete inhibition of ricin-induced morphological changes, while the use of the monoclonal antibodies (either directed against subunit A or subunit B) induced only a partial inhibition of cell confluence and phase-shift and complete inhibition of all other parameters. These variations may originate from the fact the polyclonal antibodies are composed of antibodies that target different epitopes simultaneously, increasing the neutralization effect.
The DHM in our assay was able to detect single cells for analysis only at confluence levels of under 70%. Above this level, an inconsistency in consecutive frame analyze was observed, hampering the analysis after 20 h of incubation (due to cell replication). In order to detect changes in HeLa and Vero cells with a low dose of toxins—less than 0.3 ng/mL—prolonged intoxication times are crucial. To that end, cells should be initially cultured at less than 10% confluence. Another option, not tested in this study, is to limit cell replication (by serum starvation, etc.).
The characterization of toxin-mediated morphology changes of single cells, detected by DHM, was also examined with shiga toxin [45
], resulting in cell rounding and increased cell thickness. Interestingly, morphological changes observed upon etoposide intoxication occur prior to the reduction in cell metabolic activity or viability as measured by MTS assay [46
]. In addition, monitoring cell cycle arrest [35
], and cytotoxicity assessment [31
] by DHM were found comparable to established assays such as flow cytometry, fluorescence-based methods, and trypan blue assays.
Phototoxicity has the ability to increase the intoxication effect of ricin, as was demonstrated by Song et al. [47
] using fluorescent microscopy for cell imaging, thus raising a question regarding similar effects in the DHM setup. The phototoxicity level in DHM was addressed by several groups who have demonstrated that the light intensity is six order of magnitude lower than intensities typically associated with confocal fluorescent microscopy and well below the phototoxicity levels [37
]. The advantage of DHM is that it is a label-free technique that does not requires the additions of dyes that by interacting with the light might exacerbate the toxin effects. Additionally, we have demonstrated that by adding toxin-neutralizing antibodies we have completely inhibited all morphological and optical effects, thus implying that phototoxicity-induced nonspecific effects following ricin toxicity are minimal if any.
Cell based assays for ricin toxicity measure cell survival as the intoxication end point, usually within 48 to 72 h postexposure [16
]. As ricin-mediated protein synthesis arrest is an early cellular event, it can be detected in vitro by measuring marker protein levels 3–24 h postintoxication [6
]. In this study, we demonstrate that significant changes in cellular optical and morphology parameters occur as early as 4–6 and 2–3 h using 10 and 100 ng/mL of ricin, respectively. These results are well correlated to previous reports that measured inhibition of intracellular protein synthesis using the same toxin concentrations, and therefore strengthen the use of DHM to study and follow ricin activity. Moreover, the use of specific antitoxin neutralizing antibodies to inhibit the toxic effects on the cells is crucial in every cell-based as well as noncell-based assay. This an internal control for the specificity of the assays. When compared to other cell-based methods, DHM does not requires any additional steps to evaluate the toxin effect, thus making this method simple yet sensitive.
Overall, our results emphasize the benefits of using DHM for the detection of ricin and abrin activity in cell lines as a rapid and sensitive assay. Application of cell morphology detection in high throughput screening is used by researchers in academia and pharma in order to address cell death determination in drug discovery [49
], adverse cellular effects [51
], venom exploration [52
], and is also in common use in the clinical setting for the detection of C. difficile
The ability of DHM technology to detect early changes in cellular features upon toxin-mediated cell death processes makes it suitable for toxin activity assays, and is comparable, if not superior, to established methods.
4. Materials and Methods
4.1. Toxins and Antibodies
Purified ricin and abrin were prepared as described previously [54
]. Anti-ricin subunit A (MH1) and anti-subunit B (MH75) in-house monoclonal antibodies utilized in this study were produced from nonhuman primates [40
]. In-house anti-ricin and anti-abrin polyclonal antibodies were obtained from immunized rabbit serum [54
4.2. Cell Lines
HeLa and Vero cell lines were obtained from the American Type Culture Collection (ATCC). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (v/v) fetal bovine serum and 1% L-glutamine (200 nM) and incubated at 37 °C in 5% CO2 humidified incubator.
4.3. Digital Holographic Microscopy
Digital holographic microscopy was performed using a Holomonitor M4 microscope (Phase Holographic Imaging AB, Lund, Sweden). The Holomonitor M4 is a label-free cell analyzer for adherent cells equipped with X20 magnification objective (NA-0.4), a 0.8 mW HeNe laser (633 nm, 100 µW/cm2
, exposure time 5 ms), and a motorized stage. Digital holographic microscopy, as a quantitative phase-contrast imaging method, is a form of optical interferometry, which detects the phase delay related to the light passing through the tested object. Images were converted from wavelength interaction to cells’ representation by a computer algorithm (Hstudio 2.6, Phase Holographic Imaging AB, Lund, Sweden). This computation enables to obtain information on various morphological and optical parameters. Two different threshold settings prior to segmentation of cells were used for analyzing the experiments: Otsu thresholding (OT) [56
] and minimum error thresholding (MET) [57
]. While the former results in a more accurate identification of cells, by excluding focus debris, MET gave a more accurate cell outline. In our study, the same trend was observed using the two threshold settings and the DHM results presented in this study were analyzed according to MET threshold. In general, cell lines (1–4 × 105
) were seeded in 6-well plates at different degrees of confluence (10–30%) with or without incubation for 4 to 6 h for cell adhesion. Following incubation, cells were exposed to ricin or abrin at 1–100 ng/mL in a 3 mL volume. The neutralizing effect of anti-ricin and anti-abrin antibodies was assessed by incubating toxins with 20 μg MH1 or MH75 monoclonal antibodies or 1:100 dilution of polyclonal antibodies. Cells were monitored continuously, in time-lapse mode (every 10 min), in multiple locations in each well using a high precision motorized stage. Quantified parameters obtained from DHM in each frame were analyzed as changes relative to time zero from each area and presented as mean ± SE for all the monitored areas in each well. In this work we focused on significant changes observed during intoxication in several parameters: phase-shift (phase delay related to the light passing through the tested object), confluence, area, irregularity (determines the deviation of the cell region from the circular shape), perimeter, and roughness (indication on the smoothness of the surface of the cell). As the refractive index of the cell and the surrounding medium cannot be measured, the presented cell thickness and volume values are expressed as the optical thickness and effective-calculated volume (ECV).
4.4. Viability Assay
The cytotoxic effects of purified ricin and abrin on HeLa and Vero cells were determined by AlamarBlue viability assay (Promega, Madison, WI, USA). Briefly, cells were cultured in 96-well plates overnight at a density of approximately 104/well and then treated with 10–100 ng/mL of ricin or abrin for 3–17 h. AlamarBlue reagent was added to each well up to 10% of tissue culture medium and incubated for an additional 3 at 37 °C. Cell viability was determined by measuring absorbances at 562 nm and 600 nm.
4.5. Flow Cytometry-Apoptosis Assay
Apoptosis induction with ricin and abrin in HeLa cells was assessed by APC Annexin V/PI apoptosis kit (catalog #460932, BioLegend, San Diego, CA, USA) and analyzed using a BD Fortesa Flow Cytometer from Becton Dickinson (San Jose, CA, USA). In general, 5 × 104/well HeLa cells were cultured overnight in 24-wells plates for cell adhesion. Following cell adhesion, toxins (10–100 ng/mL) were added to cells and apoptosis measurements were performed after 4–8 h. In order to assess the effect of cell adhesion on toxin susceptibility, HeLa cells were treated with trypsin for cell detachment and washed in a fresh growth medium prior to toxin administration. Apoptosis induction was compared between the preadherent and fully adherent cells. Untreated preadherent/adhered cells were used as a control.
4.6. Scanning Electron Microscopy
HeLa cells were grown in 75 cm2 flasks at 37 °C and 5% CO2. The cells were detached using trypsin and then seeded in 24 well plates at ~50,000 cells/well on top of polylysine-coated cover slips. The cells were allowed to attach and treated with ricin for 5 h at 37 °C. After incubation, the cells were washed with phosphate-buffered saline (PBS) once and a fixative buffer containing 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M cacodylate buffer (Cat. 15960-01, EMS, Hatfield, PA, USA) was added. Cells were washed in PBS and 1% osmium tetraoxide was added for 1 h at room temperature (RT) Cells were washed in DDW and dehydration was carried out in a gradient of ethanol concentrations. After dehydration, the ethanol was substituted with CO2 in a Critical Point Drying (CPD, Ashford Kent, UK) machine and, finally, the cover slips containing cells were coated with 2 nm iridium. The cells were imaged using QuantaFEG SEM (Thermo Fisher, OR, USA) operated at 6 kV in high vacuum mode using an ET detector.
4.7. Statistical Analysis
All experiments were repeated three times and data was analyzed using GraphPad Prism5 software (San Diego, CA, USA). Results are expressed as mean ± standard error. Statistical significance was determined by 2-tailed unpaired Student’s t-test. p value ≤ 0.05 was considered to be significant.