Reagent grade chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA), Merck KGaA (Darmstadt, Germany) and Life Technologies (Carlsbad, CA, USA). Cal A was purchased from Abcam Biochemicals (Cambridge, UK) and dissolved in dimethyl sulfoxide (DMSO; SERVA Electrophoresis GmbH, Heidelberg, Germany).

UCB mononuclear cells (MNC) were extracted as previously described [27] from 3 healthy newborns after full-term pregnancies and provided by Eurocord-Slovakia, Bratislava, Slovakia. In Laboratory 1, cryopreserved UCB MNC from three donors were thawed using standard protocols and cultured in RPMI 1640-based basal medium supplemented with fetal bovine serum and antibiotics. Monocytes were removed by short-term adherence.

In Laboratory 2, UCB MNC from three independent donors obtained under identical inclusion criteria were analyzed as part of a previously published dataset [36] and used here exclusively for inter-laboratory comparison at matched irradiation doses and time points. Except for fixation timing, identical cell handling and processing steps were applied in both laboratories.

Peripheral blood (PB) MNC were isolated from six healthy adult donors by density-gradient centrifugation as previously described [39]. Monocytes were removed by adherence, yielding predominantly lymphocyte populations. Cell viability exceeded 95% in all experiments as assessed by Trypan blue exclusion.

Following equilibration, lymphocytes were divided into three treatment groups: cal A (1 nM), DMSO vehicle control (0.1%), and untreated controls. Cells were incubated on ice for approximately 40 min prior to irradiation to synchronize handling between laboratories. To avoid prolonged exposure-related toxicity, cal A and DMSO were added immediately after irradiation in both laboratories, consistent with our previous observations [36].

Cells were irradiated on ice using acute $\gamma$-rays at a dose rate of approximately 45 cGy/min. UCBL were exposed to doses of 10, 50, 100, 500, and 1000 mGy, while PBL were irradiated at 10, 100, and 1000 mGy. Irradiations were performed using ⁶⁰Co sources (Theratron Elite 100, MDS Nordion, Canada; or TERAGAM® K-02, UJP Praha, Czech Republic). Dose deviations did not exceed 5%. After irradiation, cells were rapidly warmed to 37 °C and incubated under standard culture conditions until fixation.

At defined time points post-irradiation (0.5, 1, and 2 h for UCBL; 2, 4, and 6 h for PBL), cells were cytocentrifuged onto poly-L-lysine-coated slides, fixed with paraformaldehyde, permeabilized, and blocked according to standard protocols. $\gamma$H2AX and 53BP1 were detected using well-characterized primary antibodies provided by Novus Biologicals, Abingdon, UK - monoclonal $\gamma$H2AX anti-mouse (1:800, NB100-78356) and polyclonal 53BP1 anti-rabbit (1:1000, NB100-304); followed by Alexa Fluor-conjugated secondary antibodies (Invitrogen Molecular Probes, Life Technologies, OR, USA) - 488 IgG (H+L) polyclonal anti-rabbit (1:200, A11029) and 555 IgG (H+L) monoclonal anti-mouse (1:200, 21429). Nuclei were counterstained with DAPI-containing antifade medium.

Image acquisition was performed in both laboratories using Metafer Slide Scanning Systems V 3.9 (MetaSystems GmbH, Altlussheim, Germany) coupled to Zeiss Axio Imager fluorescence microscopes (Carl Zeiss Microscopy GmbH, Gottingen, Germany) under identical acquisition settings. For each condition, at least 600 cells were analyzed. Automated scoring of $\gamma$H2AX and 53BP1 foci was conducted as previously described [16, 40] and verified by manual inspection. Co-localized foci were defined by spatial overlap within 0.5 µm. $\gamma$H2AX pan-nuclear staining was quantified in UCBL as an apoptosis-associated endpoint but was not observed in PBL.

The yields of foci (Y) were expressed per cell. At each experimental dose point in cells of each blood donor the foci-per-cell distribution was checked for compliance with the Poisson model by $\chi$² test. When individual data were pooled, standard errors (SE), standard deviations (SD) or $\pm$0.95 confidence intervals (CI) for the mean foci yield were calculated from the dispersion of donors’ individual values within the group. The group data were analyzed using factorial analysis of variance (ANOVA) together with control for the false discovery rate (FDR). For pooled datasets fitting a normal distribution and involving a single categorical independent variable, comparisons between groups were performed by one-way ANOVA, followed by Fisher’s LSD post hoc test. Associations between groups were assessed using the Spearman rank-order correlation test (ROC), while correlations between independent groups were evaluated with the Mann-Whitney U test. These statistical analyses were performed with Statistica 8.0 software (StatSoft Inc., Tulsa, OK, USA). Differences were considered statistically significant at p 0.05.

Dose-response relationships for the foci yield were fitted by the iteratively reweighted least squares method to the linear model:

$$Y = c + \alpha \times D,$$

where Y - foci yield per cell; c - the background foci level; $\alpha$ - linear coefficient representing a foci induction rate per unit dose; D - radiation dose in Gy. Goodness of fit was assessed by a linear correlation coefficient (r). The coefficient $\alpha$ was compared between blood donors or laboratories as normally distributed parameter, using the unpaired Student’s t-test. The statistical software package Dose Estimate_v.5.1 [41] was used to perform foci distribution testing and dose-response curve fitting; the latter was carried out assuming Poisson foci-per-cell distribution for calculations of coefficient errors.

For each UCBL donor, the foci-per-cell distributions in unirradiated and irradiated samples were consistent with Poisson statistics, as expected for homogeneous exposure to low-LET radiation (data not shown). Donor-specific regression coefficients are summarized in Supplementary Table 1, and pooled regressions for each laboratory are shown in Table 1 and Fig. 1.

Fig. 1.

Linear dose-response relationships for $\gamma$H2AX, 53BP1, and $\gamma$H2AX/53BP1 co-localized foci in umbilical cord blood lymphocytes (UCBL). Mean foci yields per cell ($\pm$ SE) are shown as a function of $\gamma$-irradiation dose for Laboratory 1 (cells fixed 0.5 h post-irradiation) and Laboratory 2 (cells fixed 2 h post-irradiation). Solid lines represent linear regressions fitted according to Y = c + $\alpha$ $\times$D, with corresponding regression equations and coefficients of determination (R²) displayed in each panel. Regression coefficients ($\alpha$) and correlation parameters are summarized in Table 1 and Supplementary Table 1. Differences in absolute foci yields between laboratories reflect distinct post-irradiation fixation times and donor sets and are not interpreted as direct inter-laboratory discrepancies. In all the data, the comma is used as the decimal separator. $\gamma$H2AX, phosphorylated histone H2AX; 53BP1, p53-binding protein 1.

Across both laboratories, $\gamma$H2AX, 53BP1, and $\gamma$H2AX/53BP1 co-localized foci displayed strong linear dose-response relationships within the investigated dose range, with excellent goodness-of-fit to the linear model (r $\ge$ 0.94 for all endpoints; Table 1). In Laboratory 1 (UCBL fixed 0.5 h post-irradiation), all three endpoints were significantly elevated already at 50 mGy relative to unirradiated controls, indicating early sensitivity above endogenous foci levels. In Laboratory 2 (UCBL fixed 2 h post-irradiation), co-localized foci remained the most sensitive endpoint (significant induction at 50 mGy), while separate $\gamma$H2AX and 53BP1 foci reached statistical significance at 100 mGy. This time-dependent shift in the apparent sensitivity pattern is consistent with the known differences in foci kinetics during the first post-exposure hours.

Importantly, because the two laboratories quantified foci at different post-irradiation times (0.5 h vs 2 h), absolute foci yields between laboratories are not interpreted as direct head-to-head “inter-laboratory disagreement”. Instead, the inter-laboratory comparison is focused on (i) the reproducibility of the dose-response form (linearity), (ii) the similarity of regression slopes ($\alpha$), and (iii) the consistency of conclusions regarding cal A effects under harmonized processing and scoring workflows. In this context, pooled regression analysis indicated highly comparable $\gamma$H2AX induction rates per unit dose between laboratories, whereas $\alpha$ values for 53BP1 and co-localized foci were numerically lower in Laboratory 2; these differences did not reach statistical significance (Table 1). Overall, the narrow range of $\alpha$ values supports reproducible dose-response relationships despite differences in fixation time and donor sets.

Together, these results establish that the $\gamma$H2AX/53BP1 IRIF assay provides robust, quantitative, and highly sensitive detection of radiation-induced DNA damage in UCBL within the low-dose range, and that dose-response relationships remain reproducible across independent laboratories when protocols are workflow-aligned and key steps are harmonized.

Laboratory 1 assessed the effects of irradiation dose (0, 10, 50, 100 mGy), treatment condition (cal A, DMSO vehicle, untreated), and post-irradiation time (0.5, 1, 2 h) on $\gamma$H2AX, 53BP1, and co-localized foci, as well as on the frequency of $\gamma$H2AX pan-nuclear staining in UCBL (Table 2). Multivariate ANOVA identified irradiation dose as the only significant factor across foci endpoints. Treatment condition and time within the first 2 h did not reach statistical significance, indicating that foci levels were primarily determined by dose under the applied experimental conditions.

Univariate testing confirmed a highly significant dose dependence for $\gamma$H2AX, 53BP1, and co-localized foci (all FDR-adjusted p 0.00001; Table 2). In contrast, neither treatment (cal A or DMSO) nor post-irradiation time significantly altered foci yields. $\gamma$H2AX pan-staining showed no significant dependence on dose, treatment, or time within this experimental window (Table 2), supporting its interpretation as a distinct endpoint not tracking with repair foci formation under low-dose conditions.

Vehicle Control and Cal A Effect in Laboratory 1

Untreated UCBL were compared with matched DMSO-treated samples at identical doses and time points. No significant differences were detected, validating DMSO as an appropriate vehicle control. For clarity, untreated data were omitted from Fig. 2, which focuses on direct cal A vs DMSO comparisons. As shown in Fig. 2, foci yields are presented across multiple doses, time points, and treatments to illustrate overall assay behavior under the experimental design; no direct quantitative comparison between post-irradiation time points is intended.

Fig. 2.

Dose- and time-dependent induction of DNA repair foci in UCBL treated with DMSO or cal A - Laboratory 1. Panels show mean $\pm$0.95 confidence interval from 3 experiments for (A) $\gamma$H2AX foci per cell, (B) 53BP1 foci per cell, (C) co-localized $\gamma$H2AX/53BP1 foci per cell, and (D) percentage of $\gamma$H2AX pan-stained cells. For each endpoint, cells were irradiated with 0, 10, 50, and 100 mGy and analyzed at 0.5, 1, and 2 h post-irradiation/treatment. Concentrations - DMSO (0.1%) and cal A (1 nM). A minimum of 600 nuclei per condition was evaluated. Asterisks indicate significance versus the unirradiated control (0 mGy) within the same treatment group (DMSO or cal A) as determined by Fisher’s LSD post hoc testing: * p 0.05, ** p 0.01, *** p 0.001 (statistical procedures as described in Methods). Note: the figure illustrates the multivariate experimental design (dose $\times$ time $\times$ treatment) and endpoint behavior across conditions. It is intended to visualize qualitative trends and experimental structure rather than to enable direct quantitative comparison between different post-irradiation time points. In all the data, the comma is used as the decimal separator.

As shown in Fig. 2, $\gamma$H2AX, 53BP1, and co-localized foci increased with irradiation dose at all time points (0.5, 1, 2 h). Critically, cal A at 1 nM did not produce a systematic increase in foci yield or altered kinetics relative to DMSO controls at any dose or time point examined. $\gamma$H2AX pan-staining remained low and comparable between treatments. Strong positive correlations were observed among $\gamma$H2AX, 53BP1, and co-localized foci, whereas pan-staining did not correlate with foci-based endpoints, consistent with an apoptosis-associated signal distinct from DSB repair foci formation. Representative immunofluorescence images are provided in Fig. 3.

Fig. 3.

Representative immunofluorescence images of UCBL nuclei analyzed for $\gamma$H2AX/53BP1 foci, irradiated with 0, 10, 50, or 100 mGy and fixed at 0.5 h post-irradiation, followed by treatment with cal A (1 nM). Shown are examples of (A) nuclei exhibiting discrete $\gamma$H2AX foci (red), 53BP1 foci (green), co-localized $\gamma$H2AX/53BP1 foci (yellow), and (B) $\gamma$H2AX pan-staining (diffuse nuclear red staining). Scale bar = 10 µm. Nuclei were counterstained with DAPI (blue). The images represent typical results from more than 600 cells scored per condition. Automated image acquisition was performed using the Metafer Slide Scanning System and Zeiss Axio Imager Z2 microscope under identical settings across all samples. These representative micrographs illustrate the main categories of DNA damage response endpoints evaluated quantitatively in Fig. 2.

To enable a time-matched inter-laboratory comparison, we evaluated UCBL data acquired at 2 h post-irradiation (Laboratory 1:2 h time point from the present study; Laboratory 2:2 h dataset reported previously and included here specifically for comparative evaluation). A three-way factorial ANOVA (Laboratory, Treatment, Dose) was performed (Table 3).

A significant laboratory factor was detected for $\gamma$H2AX, 53BP1, and co-localized foci (all p 0.00001), indicating systematic differences in absolute foci yields between laboratories. In contrast, no laboratory effect was observed for pan-staining (p = 0.234). The treatment factor (1 nM cal A vs 0.1% DMSO) was not significant for any foci endpoint (all p $>$ 0.3). For pan-staining, the treatment effect showed a near-threshold trend (p = 0.078) that did not remain significant after correction, suggesting at most a weak influence under these conditions. Irradiation dose remained the dominant determinant of foci induction (all foci endpoints p 0.00001 after FDR adjustment), while pan-staining again showed no dose dependence (p = 0.546).

Fig. 4 visualizes the 2 h comparison and confirms three core outcomes: (i) robust dose-dependent induction of $\gamma$H2AX, 53BP1, and co-localized foci in both laboratories, (ii) systematic inter-laboratory offsets in absolute foci yields, and (iii) the absence of a detectable stabilizing effect of 1 nM cal A relative to DMSO. Correlation analyses supported high internal consistency among the three foci endpoints in each laboratory, while pan-staining remained uncorrelated. Non-parametric Mann-Whitney testing confirmed significant inter-laboratory differences in absolute foci counts for all three foci endpoints, but not for pan-staining.

Fig. 4.

Inter-laboratory comparison of $\gamma$H2AX, 53BP1, co-localized $\gamma$H2AX/53BP1 foci, and $\gamma$H2AX pan-staining 2 h after low-dose $\gamma$-irradiation. (A) Mean $\pm$ SD foci per cell for $\gamma$H2AX (red), 53BP1 (green), and $\gamma$H2AX/53BP1 co-localized foci (yellow) at 0, 10, 50, and 100 mGy. (B) Mean $\pm$ SD percentage of $\gamma$H2AX pan-stained nuclei under the same conditions. Data are shown separately for Laboratory 1 and Laboratory 2, each analyzing three independent donors. Cells were treated immediately after irradiation with either cal A (1 nM) or control (0.1% DMSO). At least 600 nuclei per condition were evaluated using automated image acquisition and scoring. Asterisks indicate statistically significant differences between Laboratory 1 and Laboratory 2 under identical irradiation and treatment conditions, as determined by Fisher’s LSD post hoc testing: * p 0.05, ** p 0.01, *** p 0.001. In all the data, the comma is used as the decimal separator.

Taken together, the time-matched comparison shows that while absolute yields can differ between laboratories, dose-response relationships and the qualitative treatment conclusion (no cal A stabilization at 1 nM) are reproducible under harmonized workflows.

To assess whether findings in cryopreserved UCBL generalize to fresh adult lymphocytes and to reduce potential freeze-thaw confounding, peripheral blood lymphocytes (PBL) were analyzed in Laboratory 2 across doses and time points (Table 4; Fig. 5). Factorial ANOVA demonstrated highly significant dose effects for all foci endpoints at 2 h (FDR-adjusted p 0.001) and similarly when including extended time points (2, 4, 6 h) at 0 and 100 mGy (FDR-adjusted p 0.005). Treatment and post-irradiation time did not significantly affect foci levels (FDR-adjusted p $>$ 0.3 and p $>$ 0.4, respectively).

Fig. 5.

Induction of DNA repair foci in PBL across treatments and irradiation conditions - Laboratory 2. (A) Mean numbers of $\gamma$H2AX, 53BP1, and co-localized $\gamma$H2AX/53BP1 foci per cell [$\pm$0.95 confidence interval (CI) from 3 experiments] are shown for untreated irradiated cells, DMSO-treated cells, and cal A (1 nM)-treated cells at 0, 10, and 1000 mGy, assessed 2 h post-irradiation/treatment. (B) Mean $\gamma$H2AX foci per cell ($\pm$0.95 CI) at 0 and 100 mGy measured at 2, 4, and 6 h post-irradiation/treatment. (C) 53BP1 foci per cell ($\pm$0.95 CI) under the same conditions. (D) $\gamma$H2AX/53BP1 co-localized foci per cell ($\pm$0.95 CI) under the same conditions. A significant reduction of $\gamma$H2AX/53BP1 co-localization in-between treatments as determined by Fisher’s LSD post hoc testing is indicated by ** p 0.01. A minimum of 600 nuclei were analyzed per condition using automated image acquisition and scoring. In all the data, the comma is used as the decimal separator.

Notably, $\gamma$H2AX pan-nuclear staining was not detected in PBL under any tested condition, in contrast to UCBL where pan-staining was consistently measurable. This difference indicates that pan-staining is not an inherent feature of lymphocytes under low-dose irradiation but depends on cell system and/or pre-analytical handling such as cryopreservation.

Under comparable experimental conditions, both UCBL (in both laboratories) and PBL exhibited dose-dependent induction of $\gamma$H2AX, 53BP1, and co-localized foci. However, two consistent differences were observed. First, $\gamma$H2AX pan-staining was present in UCBL but absent in PBL, suggesting that apoptotic $\gamma$H2AX signaling is influenced by cell type and/or cryopreservation-related pre-analytical factors. Second, cal A at 1 nM did not produce stabilizing effects in either system across the low-dose range; the only statistically significant treatment-related effect was a reduction of co-localized foci in PBL at 1000 mGy.

Overall, these results support high robustness of IRIF dose-response relationships under harmonized workflows while indicating that the impact of cal A is strongly context-dependent and not reliably stabilizing under the tested conditions.

Across two independent laboratories, $\gamma$H2AX, 53BP1, and $\gamma$H2AX/53BP1 co-localized foci in UCBL showed reproducible linear dose-response relationships under harmonized low-dose workflows. $\gamma$H2AX pan-nuclear staining represented a biologically distinct endpoint, uncoupled from dose-response behavior and repair foci formation under the examined low-dose conditions. Importantly, under harmonized low-dose conditions, cal A (1 nM) did not provide any measurable stabilization or enhancement of IRIF yields. Where data are shown across multiple post-irradiation times, comparisons are interpreted qualitatively and within each time point, rather than as direct kinetic analyses.

Although absolute IRIF yields differed between laboratories, the form and linearity of the dose-response relationships were preserved across endpoints. Crucially, the qualitative conclusion regarding cal A was consistent across laboratories. Although these findings confirm the expected low-dose sensitivity of $\gamma$H2AX/53BP1 foci induction, the principal contribution of this study lies in demonstrating that harmonized experimental workflows yield reproducible dose-response relationships across independent laboratories. These findings show that inter-laboratory reproducibility of IRIF-based biodosimetry depends primarily on workflow harmonization rather than on pharmacological modulation of DNA damage signaling.

Analysis of freshly isolated PBL served as a within-laboratory control to assess whether conclusions derived from cryopreserved UCBL were workflow-dependent. PBL showed consistent dose-dependent induction of all IRIF endpoints, with foci yields at 100 mGy exceeding those observed in UCBL from Laboratory 1 and closely matching UCBL data from Laboratory 2. Importantly, $\gamma$H2AX pan-staining was absent in PBL under all experimental conditions, underscoring system-dependent differences between UCBL and PBL under the examined conditions and minimizing apoptosis-related confounding in fresh cells. At the highest dose tested (1000 mGy), cal A treatment in PBL was associated with a reduction in $\gamma$H2AX/53BP1 co-localized foci relative to DMSO controls, a pattern consistent with altered repair kinetics or apoptosis-related chromatin changes rather than foci stabilization. Under the tested low-dose, early post-irradiation conditions, this study rules out cal A as a generally applicable strategy for improving IRIF-based biodosimetric robustness.

Phosphatase inhibition by cal A has been reported to affect $\gamma$H2AX signaling in several experimental systems, although with substantial context dependence [32, 34]. In adult PBL, Kuefner et al. (2013) [31] reported prolonged $\gamma$H2AX foci persistence after low-dose irradiation in the presence of cal A, while effects on 53BP1 were limited. Importantly, these experiments differed from the present study in post-irradiation timing, workflow structure, and analytical endpoints. By contrast, under harmonized low-dose workflows optimized for early biodosimetric readout, the present study did not detect a consistent stabilizing effect of calyculin A on $\gamma$H2AX or 53BP1 foci. These discrepancies are more plausibly explained by differences in workflow design, post-irradiation timing, dose range, and analytical thresholds than by cal A efficacy per se. Recent reviews emphasize that such context dependency is a central challenge in biodosimetry assay standardization [17, 42].

The PBL dataset was included to test the internal consistency of conclusions under conditions free of cryopreservation-related stress. While UCBL showed no cal A effect across the examined dose range, PBL displayed a reduction in $\gamma$H2AX/53BP1 co-localization at high dose under cal A treatment. This observation supports the notion that cal A’s impact depends not only on timing and workflow but also on intrinsic biological properties of the cell system. The results do not contradict earlier reports of cal A effects, but delimit their applicability: under rapid, harmonized low-dose biodosimetry workflows, pharmacological phosphatase inhibition does not improve assay robustness.

Systematic offsets in absolute IRIF yields between laboratories are a well-recognized feature of multicenter biodosimetry studies and do not, by themselves, indicate methodological failure [15, 17, 46]. In the present study, such offsets did not alter dose-response linearity or qualitative treatment conclusions. These findings are directly relevant to ongoing biodosimetry standardization efforts within international networks such as RENEB and EURADOS, which increasingly prioritize workflow harmonization over absolute inter-laboratory yield matching. Accordingly, the objective of inter-laboratory comparison in this study was not to eliminate absolute yield differences, but to determine whether harmonized workflows preserve interpretable and reproducible dose-response behavior. Recent methodological analyses converge on the same conclusion: reproducibility in IRIF-based biodosimetry is achieved through disciplined workflow alignment rather than through incremental assay sensitivity gains [47, 48].

Consistent correlations among $\gamma$H2AX, 53BP1, and co-localized foci support co-localization as the most robust readout for low-dose DSB detection under harmonized assay conditions. By contrast, $\gamma$H2AX pan-nuclear staining did not track with IRIF-based endpoints and should be interpreted separately from repair-associated foci under low-dose conditions [16, 26, 27, 37, 38]. This distinction is particularly relevant when evaluating cal A, as phosphatase inhibition can alter chromatin-associated signaling without improving repair foci-based readouts.

Standardized IRIF assays without pharmacological modifiers provide reliable and interpretable dose-response information in the low-dose range relevant for biodosimetry [2, 6, 49, 50]. While cal A remains useful in specific radiobiological applications, such as premature chromosome condensation assays [51, 52, 53, 54], its routine use as an enhancer of IRIF-based biodosimetry is not supported under harmonized low-dose workflows.

Future work should focus on validating IRIF-based biodosimetry under realistic operational constraints, including delayed processing and inter-operator variability, rather than on pharmacological signal amplification [55]. From a translational perspective, these findings argue for abandoning cal A as a default ‘fix’ for inter-laboratory variability and instead prioritizing disciplined workflow harmonization.