A prospective study on histone ?-H2AX and 53BP1 foci expression in rectal carcinoma patients: correlation with radiation therapy-induced outcome

DNA damage and its repair were evaluated up to 24 h after exposure to 0.5 Gy or 2 Gy
of X-rays in vitro or after 5 first clinical radiation fractions. The extent of DNA damage was measured
by counting the number of histone ?-H2AX foci, a sensitive marker of DNA DSBs 30]. The mean data from 500 nuclei were determined for the cell samples from each tested
individual (Fig. 1). The means for each tested group of individuals are also shown in Fig. 1.

Fig. 1. Comparison of histone ?-H2AX foci in PBMCs derived from control donors and unselected
RC patients. a DNA damage assessed by means of the histone ?-H2AX assay in non-irradiated and b–d in irradiated PBMCs derived from unselected RC patients (triangles), as compared
to the cells from apparently healthy donors (circles). Initial (b), residual (c – 0.5 Gy, 24 h, d – 2 Gy, 24 h) DNA damage were assessed in PBMCs after irradiation with 0.5 Gy (b, c) or 2 Gy (d) in vitro. Filled squares represent the mean values (± SE) for the respective group

The parameters on initial, residual and baseline DNA damage assessed by histone ?-H2AX
for each individual, as well as age, sex, and grade of GI toxicity after RT are given
Fig. 1 and in Additional file 1: Table S3. Although non-irradiated cells of some RC patients showed remarkably lower
intrinsic DNA damage, i.e. in the range of controls, the mean value of background DNA damage (Fig. 1a) was significantly (p??0.005) higher (0.5?±?0.1 foci/nucleus) in the group of unselected RC patients,
as compared to the group of healthy controls (0.1?±?0.03). Likewise, irradiated in vitro blood lymphocytes showed higher (p??0.005) initial (Fig. 1b, 0. 5 Gy, 30 min) and residual (p??0.005, Fig. 1c and d, 0.5 Gy and 2 Gy, 24 h) expression of the ?-H2AX foci.

In addition, the foci numbers of 53BP1, a sensor of DNA damage 26], were compared between 10 healthy controls and 47 RC patients. As seen in Additional
file 1: Figure S1 and Table S4, the mean background expression levels of 53BP1 (Additional
file 1: Figure S1A) were very similar in two groups. However, the mean expression of radiation-induced
53BP1 foci (Additional file 1: Figure S1, part B) was not significantly higher (3.6?±?1.8 foci/nucleus) in the
group of RC patients than that in control group (2.4?±?0.4 foci/nucleus) probably
because of the enormous data scattering in the RC group. The numbers of residual 53BP1
foci detected 24 h post-IR (Additional file 1: Figure S1, parts C and D) were found to be significantly (p??0.05 and p??0.005 after 0.5 and 2 Gy, respectively) higher in the PBMCs derived from RC patients
than that of healthy individuals.

Next, we compared the expression of ?-H2AX and 53BP1 per one and the same nucleus
at different time post-IR and radiation doses. Judging from the correlation coefficients
given in Additional file 1: Figure S2, there was no (Additional file 1: Figure S2, part A) or weak correlation (Additional file 1: Figure S2, part B) between background (0 Gy) or radiation-induced (30 min after
irradiation with 0.5 Gy) expression of both proteins, respectively. At the same time,
a strong (R
²
?=?0.92 and R
²
?=?0.83) correlation was found between residual amounts of ?-H2AX and 53BP1 foci (Additional
file 1: Figure S2, parts C and D).

Out of 53 prospectively recruited RC patients, 27 exhibited an adverse GI reaction
to RT, including grade 2 and grade 3 according to RTOG score (see Additional file 1: Table S1). Based on the clinical GI reaction of RT patients we analyzed retrospectively the initial, residual and background DNA damage measured by histone ?-H2AX between
the groups of RC patients with normal (RTOG grade 0 and 1, n?=?26) and an adverse (RTOG grade 2 and 3, n?=?27) clinical reaction to RT (Fig. 2). As seen in Fig. 2, background, induced or residual DNA damage in PBMCs from RC patients with normal
or adverse clinical reaction was higher than that from control donors. However, there
was no difference between the both groups (grade 0–1 and 2–3) of RC patients in all
parameters studied (Fig. 2a-d). Mostly similar data were obtained with the 53BP1 foci except that there was no
difference between the background numbers of 53BP1 foci counted in all 3 groups (Additional
file 1: Figure S3, parts A-D).

Fig. 2. Histone ?-H2AX foci in PBMCs derived from control donors and normally-reacting and
radiosensitive (grade 2–3) RC patients. a DNA damage assessed by means of the histone ?-H2AX assay in non-irradiated and b–d irradiated PBMCs derived from normally-reacting RC patients (grade 0 and 1, up triangles)
and radiation-sensitive (grade 2 and 3, down triangles) cancer patients compared to
cells from apparently healthy donors (circles). Initial (b, 0.5 Gy, 30 min post-IR), residual DNA damage 24 h after in vitro 2 Gy (c) or 72 h after 5 clinical radiation fractions (d) were assessed in PBMCs after irradiation either in vitro (b, c) or in vivo (d). Filled squares represent the mean values (± SE) for the respective group. “n.s.”
indicates that the difference was not highly significant (p??0.05). “n.d.” means not determined. Clinical GI toxicity to RT was controlled at
the end of RT (see Additional file 1: Table S2) and used as an indicator for clinical radiosensitivity according to the
RTOG score 27]

Further, we split the group of patients (Fig. 3) with an adverse GI reaction to RT (grade 2 and 3) into 2 subgroups showing either
grade 2 (n?=?19) or grade 3 (n?=?8) reaction and compared DNA damage between these groups and a group of normally-reacting
(grade 0–1) RC patients. As seen in Fig. 3, we found no differences in the baseline, induced or residual DNA damage assessed
by the ?-H2AX foci between the groups.

Fig. 3. Histone ?-H2AX foci in PBMCs derived from normally-reacting and radiosensitive (grade
2 and grade 3) RC patients. a DNA damage assessed by means of the histone ?-H2AX assay in non-irradiated and b–d irradiated PBMCs derived from normally-reacting RC patients (grade 0 and 1, up triangles)
compared to cells from radiation-sensitive (GI toxicity, Additional file 1: Table S2) RC patients with grade 2 (down triangles, n?=?19) and grade 3 (up triangles, n?=?8). Peripheral lymphocytes were prepared from the blood samples derived from RC
patients. For details, see legend to Fig. 2

In addition to the irradiated in vitro cells, as mentioned in the Methods, blood samples were withdrawn from all recruited
RC patients after 5 clinical fractions. As seen in Fig. 4a, the mean number of ?-H2AX foci per patient’s sample after 5 clinical fractions was
significantly (p??0.05) higher (0.90?±?0.10) than that before RT (0.55?±?0.07). However, the amounts
of ?-H2AX foci (1.0?±?0.3) after clinical irradiation in a group of RC patients with
adverse (grade 3, n?=?8) clinical reaction to RT were similar to that of the unselected (n?=?53) RC patients.

Fig. 4. Effect of clinical radiation on the expression of histone ?-H2AX and 53BP1 foci in
blood lymphocytes. a DNA damage was assessed by means of the histone ?-H2AX and b 53BP1 assays before (up triangles) and after 5 clinical fractions in PBMCs derived
from unselected (right triangles) RC patients compared with RC patients with an adverse
(grade 3, down triangles) clinical GI reaction to RT. Filled squares represent the
mean values (± SE) for the respective group. “n.s.” indicates that the difference
was not highly significant (p??0.05)

The quantification of 53BP1 foci after 5 clinical radiation fractions (Fig. 4b) was conducted in a smaller group (n?=?46 vs. n?=?53 tested for ?-H2AX) RC patients, which however, contained almost all (n?=?7) clinically radiation sensitive RC patients with grade 3 GI reaction to RT. Comparison
of the mean number of 53BP1 foci per patient’s sample after 5 clinical fractions revealed
significantly (p??0.001) increased foci numbers after clinical irradiation (0.87?±?0.06 vs. 0.6?±?0.06 before RT) for the whole group tested. A subset of clinically irradiated
RC patients with an adverse clinical reaction to RT showed also an increased but similar
number of 53BP1 foci (0.90?±?0.13) as the group of unselected RC patients.

Next, we asked whether the tumor stage can influence the baseline, induced and residual
DNA damage in blood cells of RC patients. We compared the expression of ?-H2AX and
53BP1 foci in the blood lymphocytes of RC patients with different UICC tumor stages
(Additional file 1: Table S2). As seen in Fig. 5, no significant difference in the ?-H2AX foci numbers was observed between tumor
stage II, III or IV. However, the mean number of the background, induced or residual
amount of the ?-H2AX foci in the group with stage IV has the tendency to be always
lower than that of the group with the tumor stage III. The same tendency was observed
in case of 53BP1 foci (Additional file 1: Figure S4).

Fig. 5. Correlation between the ?-H2AX foci expression and tumor staging (II, III, IV). Peripheral
lymphocytes were prepared from the blood samples derived from RC patients. a Foci counting for ?-H2AX was performed in non-irradiated, b irradiated in vitro with 0.5 and c 2 Gy samples 30 min and 24 h post-IR or d after 5 clinical fractions. Filled squares represent the mean values (± SE) for the
respective group. Locoregional tumor stage was evaluated according to the standard
UICC criteria (endoscopy, endorectal ultrasound and MRI) which gave 11, 35, and 7
cases (Additional file 1: Tables S1 and S2, pre-RT) scored as stage II, III, and IV, respectively. “n.s.”
indicates that the difference was not highly significant (p??0.05)

In addition, we analyzed if the TRG (Additional file 1: Table S2) after curative RT can be predicted on the basis of both protein markers
(Fig. 6). Thus we compared the groups with “bad” (TRG 0–2, n?=?34) and “good” (TRG 3–4, n?=?19) response to RT. However, we found no differences in the background, induced
or residual (in vitro and in vivo) ?-H2AX foci between both groups (Fig. 6). Likewise, no difference between the groups with “bad” (TRG 0–2) and “good” (TRG
3–4) response to RT was observed in the degree of the induction of DNA damage (Additional
file 1: Figure S5).

Fig. 6. Correlation between the ?-H2AX foci expression and tumor regression grade (TRG). DNA
damage assessed by means of the ?-H2AX foci expression in non-irradiated and irradiated
peripheral lymphocytes of RC patients with different tumor regression grade (TRG,
Additional file 1: Table S2). Up and down triangles show ?-H2AX foci amounts in the cells of RC patients
with TRG 0–2 and TRG 3–4, respectively. Filled squares represent the mean values (±
SE) for the respective group. “n.s.” indicates that the difference was not highly
significant (p??0.05)