|Home | About | Journals | Submit | Contact Us | Français|
Peripheral blood lymphocytes of patients with hematological malignancies or solid tumors, such as renal cell carcinoma or prostate cancer, display epigenetic aberrations (loss of synchronous replication of allelic counterparts) and genetic changes (aneuploidy) characteristic of the cancerous phenotype. This study sought to determine whether such alterations could differentiate breast cancer patients from cancer-free subjects.
The HER2 locus—an oncogene assigned to chromosome 17 whose amplification is associated with breast cancer (BCA)—and the pericentromeric satellite sequence of chromosome 17 (CEN17) were used for replication timing assessments. Aneuploidy was monitored by enumerating the copy numbers of chromosome 17. Replication timing and aneuploidy were detected cytogenetically using fluorescence in situ hybridization technology applied to phytohemagglutinin-stimulated lymphocytes of 20 women with BCA and 10 control subjects.
We showed that both the HER2 and CEN17 loci in the stimulated BCA lymphocytes altered their characteristic pattern of synchronous replication and exhibited asynchronicity. In addition, there was an increase in chromosome 17 aneuploidy. The frequency of cells displaying asynchronous replication in the patients' samples was significantly higher (P < 10-12 for HER2 and P < 10-6 for CEN17) than the corresponding values in the control samples. Similarly, aneuploidy in patients' cells was significantly higher (P < 10-9) than that in the controls.
The HER2 and CEN17 aberrant replication differentiated clearly between BCA patients and control subjects. Thus, monitoring the replication of these genes offers potential blood markers for the detection and monitoring of breast cancer.
A dominant aspect of the epigenetic profile of a DNA segment is the timing of its synthesis during the S-phase of cell division [1–3]. Usually, active loci replicate earlier than silent ones [1–4]. An archetypal example of the close correlation between replication timing and epigenetic activity, first recognized nearly 50 years ago, is the delayed replication of the inactive X chromosomes in mammalian female cells when compared with their active X counterparts [5,6]. A similar correlation is seen in the allelic counterparts of monoallelically expressed genes, which replicate asynchronously, whereas the allelic counterparts of biallelically expressed genes replicate synchronously [7–9].
A simple method for evaluating the temporal order of allelic replication uses the fluorescent in situ hybridization (FISH) assay [10,11]. Using FISH, it has been demonstrated that the expressed counterpart of a monoallelic gene usually replicates earlier than the inactive counterpart [11–15]. Of major importance for this work is the fact that the temporal order of replication of two allelic counterparts (either synchronous or asynchronous) is preserved even in cells in which these alleles are not expressed . Therefore, the order of replication can be used as an epigenetic marker indicating whether a particular gene is programmed to express biallelically or monoallelically [16–19].
Because replication markers are displayed even in tissues where the gene itself is not expressed, researchers have been able to study replication patterns of a large repertoire of genes in peripheral blood lymphocytes after application (in vitro) of phytohemagglutinin (PHA), a stimulator of T-lymphocyte division. Investigators have thereby established the asynchronous allelic replication of various human monoallelically expressed genes, such as those subjected to X inactivation [20,21], as well as numerous imprinted genes [12–15,22–24]. In addition, PHA-stimulated lymphocytes have revealed the synchronous replication of various biallelically expressed genes involved in generating or limiting malignant growth (the so-called “cancer genes”), including tumor suppressor genes TP53 and RB1 and oncogenes C-MYC and HER2 [22–26].
Yet, in the peripheral blood lymphocytes of patients with hematological malignancies  or solid tumors—such as renal cell carcinoma  or prostate cancer [23,24]—genes that normally replicate synchronously (as exemplified by the above-noted “cancer genes”) lose their synchronous mode and replicate asynchronously.
Moreover, certain chromosome-specific sequences (pericentromeric, unexpressed DNA arrays), which normally display synchronous replication of their homologous counterparts, can also exhibit asynchronicity in lymphocytes of patients with various types of cancer, including ovarian , hematological , and prostate [23,24] malignancies. The aberrant replication patterns of the pericentromeric arrays seem to be associated with chromosomal malsegregation, which engenders increased levels of sporadic aneuploidy in the patients' blood cells [23,26,27].
Our present study aimed to determine whether replication markers in peripheral blood lymphocytes can differentiate between breast cancer patients (BCA) and healthy controls.
We show here that in PHA-stimulated lymphocytes of BCA patients, the HER2 oncogene locus, assigned to the long arm of chromosome 17 (17q11.2–12) , fails to maintain its characteristic synchronous mode of allelic replication and replicates asynchronously. In addition, the pericentromeric array of chromosome 17 (CEN17), which normally exhibits a synchronous mode of allelic replication, reveals a loss of synchrony accompanied by chromosome 17 aneuploidy.
Peripheral blood samples were obtained from 20 consenting women diagnosed with breast cancer (BCA) and from 10 consenting cancer-free control (CON) women. Nine of the 20 BCA patients donated samples “at diagnosis” (before any therapeutic modality) and 11 donated samples during the course of chemotherapy (“at treatment”). The 20BCA patients ranged in age from 40 to 79 years, with a mean ± SD of 55 ± 12 years. The two subgroups of BCA patients did not differ significantly in age (P > .70): the first, which donated at diagnosis, ranged from 29 to 79 years; and the second, which donated at treatment, ranged from 42 to 73 years (means ± SDs of 56 ± 15 and 54 ± 9 years, respectively). The ages of CON subjects ranged from 20 to 66 years, with a mean ± SD of 43 ± 16 years. The age differences between the BCA and CON groups were insignificant (P > .05).
Two of the 30 women in the study carried a mutated BRCA1 gene. These mutation carriers belonged to the BCA group—one to the patients analyzed “at diagnosis” and the other to the group sampled “at treatment.”
Each sample (0.2 ml of whole blood after heparinization) was transferred, for short-term culturing, into 5 ml of F10 medium supplemented with 20% fetal calf serum, 3% PHA, 0.2% heparin, and 1% penicillin/streptomycin (Biologic Industries, Beit-Haemek, Israel). Cultures were incubated at 37°C for 72 hours; colchicine (Sigma, St Louis, MO) was added to a final concentration of 0.1 µg/ml for 1 hour, followed by hypotonic treatment (0.075 M KCl at 37°C for 15 minutes) and four washes, each with a fresh, cold (-20°C) 3:1 methanol-acetic acid solution. The cell suspensions were stored at -20°C until use for FISH.
Two directly labeled commercial probes were used, each identifying a specific region of chromosome 17: HER2 for 17q11.2–q12 (32-190001; Vysis, Chicago, IL) and the α-satellite probe specific for centromere 17 (CEN17; 32-130017; Vysis). The oncogene HER2 probes is a member of the epidermal growth factor receptor or HER (erb) gene family, whose amplification and/or overexpression are highly associated with human breast cancer . The CEN17 probe, usually used for the enumeration of chromosome 17, identifies a specific noncoding repetitive pericentromeric array, whose asynchronous replication has been implicated in the appearance of aneuploidy .
Slide preparation, in situ hybridization, post washing, and detection were all performed in accordance with the previously described protocol , with slight modifications. The two probes were diluted to 1:100 (HER2) in catalog no. D003 or to 1:200 (CEN17) in catalog no. D001 (Ingen's DenHyb hybridization solutions; Insitus Biotechnologies, Albuquerque, NM). These replaced the hybridization solution supplied with the probes. Five microliters of probe solution was placed on the target area of the sample slides, which was covered with a 12-mm round silanized coverslip (Insitus Biotechnologies) and sealed with rubber cement. The slides were placed in a microheating system (True Temp; Robbins Scientific, Sunnyvale, CA) at 76°C and denatured for 6 minutes at that temperature. The True Temp was turned off, and the slides were allowed to hybridize overnight in the instrument.
Slides were analyzed blindly using an Olympus BH2 fluorescent microscope (Olympus, Hamburg, Germany) with a triple band-pass filter (Chroma Technology, Brattleboro, VT). For replication analyses, at least 100 cells exhibiting two distinct, well-defined fluorescence signals were scored from each sample and for each tested probe. Signals were divided into two categories: a single dot (singlet; S) representing a nonreplicated DNA sequence and a double dot (doublet; D) indicating a replicated sequence. Thus, cells exhibited patterns signaling synchronicity (SS cells or DD cells; Figure 1, A–D) or those signaling asynchronicity (SD cells; Figure 1, E and F). Accordingly, a high frequency of SD cells is indicative of asynchronous replication, whereas a low frequency of SD points to synchronous replication . Similarly, a high frequency of SS cells reveals late replication of the locus in question, and a high frequency of DD cells shows an early replication . For each sample and for each probe, the frequencies of SD, SS, and DD cells in the total population of cells were recorded.
For the determination of aneuploidy, at least 200 cells of each sample prepared for the CEN17 replication studies were examined. Here, the number of CEN17 fluorescent signals at interphase was recorded. Cells containing one signal were designated as cells having monosomy 17, cells with two signals were listed as normal, and cells with more than two signals were recorded as multisomy 17. From the sum of monosomic and multisomic cells in the cell population, it was possible to evaluate the frequency of aneuploidy.
The statistical significance of differences between the two populations tested was carried out using the 2-tailed Student's t test (Microsoft Excel). P values .01 or less were considered statistically significant.
The study was approved by the Tel Aviv University Institutional Helsinki Committee.
The frequency of SD lymphocytes for the HER2 and CEN17 loci was significantly higher in blood samples of the 20 women diagnosed with BCA than of the 10 women without cancer (CON) (P < 10-12 for HER2 and P < 10-6 for CEN17). The mean ± SD values for HER2 ranged from 23% to 47% (34.6% ± 4.2%) in the BCA samples versus 7% to 17% (13.6% ± 3.3%) in the CON samples. For CEN17, the values ranged from 21% to 31% (25.5% ± 3.2%) in the BCA samples versus 6% to 19% (12.8% ± 3.9%) in the CON samples (Figures 2 and and3A3A).
Yet, the BCA blood lymphocytes obtained from patients at diagnosis showed, for each locus, similar distributions of SD values to those from patients analyzed during treatment (Figure 2). The corresponding SD values of the two subgroups were comparable (P > .10 and P > .90 for HER2 and CEN17, respectively).
It is worth mentioning that the lymphocytes of the two BRCA1 mutation carriers displayed SD values (33% and 34% for HER2 and 25% and 30% for CEN17), which were within the range observed in the lymphocytes of the other BCA patients (Figure 2). Thus, the risk factor associated with a BRCA1 mutation seems to have no exacerbating effect on the level of asynchronicity in the patients' lymphocytes.
Regarding the HER2 and CEN17 loci, it should be noted that in the CON group of patients, the two SD values were similar (P > .60; Figure 3A). In this group, the two SS values (P > .60; Figure 3B) and the two DD values (P > .90; Figure 3C) were also similar, indicating that both loci replicate at the same time, probably late, as judged from the relative high SS values. However, in the BCA patients, the SD frequencies displayed by HER2 were significantly higher (P < 10-8) than those for CEN17. This was observed in nearly all the samples of this group (Figures 2 and and3A).3A). Interestingly, the high SD values obtained for HER2 in the BCA patient lymphocytes were similar (P > .05) to those we obtained for the SNRPN imprinted locus (ranging from 28% to 48%; 37.5% ± 5.4%) in a group of 25 blood samples taken from cancer-free, healthy subjects ranging in age from 20 to 82 years (data not shown). This indicates that in BCA patient lymphocytes, the SD values of HER2 reached a level observed normally for an imprinted gene.
The increased SD cell frequency seen in the BCA samples for HER2 and CEN17 was associated in both loci with a decrease in the frequency of SS cells (Figure 3B) but not that of the DD cells (Figure 3C). Accordingly, the SS cell values in the BCA samples were significantly lower (P < 10-3 for HER2 and P = 10-3 for CEN17) and the DD values were similar (P > .05 for HER2 and P > .90 for CEN17) to those in the CON samples (Figure 3, B and C). This finding points to a cancer-dependent phenomenon that accelerates the replication timing of a single allele from each studied locus but does not delay replication.
The asynchronous replication pattern of CEN17 in the BCA patient lymphocytes was accompanied by increased aneuploidy of chromosome 17 (Figure 4A). The frequencies (%) of cells displaying aneuploidy, namely, the monosomy (loss of one copy) or multisomy (gain of one or more copies), of chromosome 17 were significantly higher (P < 10-9) in the breast cancer patients than in the controls. The frequency of cells with aneuploidy (sum of monosomy and multisomy) for chromosome 17 in the BCA samples ranged between 9.5% and 19.5%, with a mean ± SD of 14.3% ± 3.2%, compared with a range of 5.0% to 8.5% (6.8% ± 1.1%) in the CON samples (Figures 4A and and5).5). The increased level of aneuploidy in the BCA patients arises mostly from monosomy and less from multisomy (Figures 4B and and5).5). The frequency of cells with monosomy for chromosome 17 in the BCA samples ranged between 8.0% and 18.0% (11.9% ± 3.0%) versus between 4.5% and 7.5% (6.0%± 1.3%) in the CON samples (Figure 4B). With respect to the monosomy values, these groups differ significantly (P < 10-7; Figure 5A). Although much lower than the monosomy frequencies, the multisomy level in the BCA samples (composed mainly of trisomy 17) was significantly higher (P < 10-3) than that of the CON samples (Figures 4B and and55).
As seen for the SD frequencies, the distribution of aneuploidy, monosomy, and multisomy values in the samples obtained from breast cancer patients at diagnosis was comparable to those obtained from patients under treatment (P > .70, P > .90, and P > .40, respectively; Figure 4).
Similarly, the BRCA1 mutation had no effect on the aneuploidy levels of the BCA patients, as it had no effect on the SD levels, with the percentage of aneuploidy, monosomy, and multisomy in the lymphocytes of the BRCA1 mutation carriers having values similar to those in the mutation-free cancer patients (Figure 4).
We investigated the temporal order of allelic replication of HER2 and a noncoding repetitive DNA sequence (CEN17) in breast cancer patients and controls. HER2 is an oncogene assigned to the long arm of chromosome 17, whose amplification and/or overexpression in breast cancers are associated with a poor outcome [28,29]. Yet, HER2 is not expressed in lymphocytes of both control subjects and BCA patients including those with HER2-amplified tumors . CEN17 is the pericentromeric DNA array of chromosome 17 and, as such, is under the control of complex epigenetic processes aimed to ensure the fidelity of chromosome segregation and genomic stability (reviewed in Gopalakrishnan et al. ). Both these loci in the peripheral blood lymphocytes of cancer-free women exhibit synchronous allelic replication. Yet, in the peripheral blood lymphocytes of BCA patients, they display a loss of synchrony and replicate asynchronously. The level of asynchrony displayed by the HER2 locus reached the levels of asynchrony normally observed for imprinted genes. These results are in line with previous studies reporting that biallelically expressed genes in PHA-stimulated lymphocytes of patients with solid tumors, such as renal cell carcinoma  or prostate cancer [23,24], lose their normal mode of synchronous replication, which characterizes biallelically expressed loci. Therefore, the loss of allelic synchronous replication in lymphocytes can be defined as a non-tumor-specific epigenetic malignant aberration. It is therefore reasonable that the mechanisms giving rise to this aberration might be observed in various other malignancies. Because the altered replication in the blood cells of prostate cancer patients was linked to enhanced methylation capacity [23,24], it is logical that the same holds true for the aberration in the blood cells of the BCA patients. This is further enforced by the fact that allelic counterparts that normally replicate asynchronously are usually differentially methylated .
We show here that the aberrant replication displayed in lymphocytes of BCA patients at diagnosis, before any pharmaceutical treatment, is similar to that observed in cells taken from the patients during chemotherapy. These results confirm a recent study showing that lymphocytes of patients with hematological malignancies demonstrate replication aberrations displayed at diagnosis, which are similar to those observed after various substantial drug treatments. However, the aberrations are eliminated after successful stem cell transplantation .
The appearance of identical replication aberrations in each of the investigated loci of all BCA, prostate cancer [23,24], and renal cell carcinoma  patients and in none of the cancer-free control subjects may speak against an inherent predisposition theory [33,34]. In our opinion, it seems to represent a phenomenon acquired later in life that is linked to the malignant transition itself.
In this respect, it is worth mentioning that the two BCA patients carrying the BRCA1 mutation displayed replication values similar to those having nonmutated BRCA1 genes. This may suggest that BRCA1 malfunctioning is not related to the BCA replication aberrations described here. Yet, two patients are not enough to conclude. Alternatively, one could speculate that the epigenetic alteration, shown here for the centromeric region of chromosome 17 and for the HER2 locus assigned to 17q12, may span further along the long arm of chromosome 17. Such a development could disturb functional properties of genes distal to HER2, namely BRCA1, which is located at 17q21 . This seems plausible on the ground that epigenetic gene alteration in cancer is not a localized event affecting discrete genes, but a generalized condition that spans broad chromosomal regions , influencing the entire genome and resulting in global epigenetic disequilibrium .
Interestingly, the loss of synchronous replication in the patients' cells was set off by the early replication of a single allele, inferred from the decrease in the SS values but not in the DD values in the BCA patients' cells. Because early replication is linked with functional genetic activity [1–6], it may correlate with the enhanced expression of oncogene HER2. However, this is not the case here because HER2 expression seems to be low in lymphocytes of BCA patients, regardless of its status in tumors . An advanced replication, not accompanied by elevated levels of expression, may be explained by the fact that replication domains can extend over large chromosomal regions containing many genes that are not necessarily regulated by the replication timing control mechanisms (discussed in Woodfine et al. ). Thus, although there is a general correlation between transcription and replication timing, the correlation between expression and replication timing often does not hold at the level of a single gene. It is reasonable to suggest, therefore, that the inherent epigenetic programming that coordinates allelic behavior turns loose (weaken) in blood lymphocytes of cancer patients. This is in agreement with previous studies indicating that the cancer-mediated abnormal asynchronous replication seems to be a global non-locus-specific phenomenon, observed in blood lymphocytes of patients with ovarian cancer , leukemia. , renal cell carcinoma , and prostate cancer [23,24]. Furthermore, the fact that asynchronous replication is also displayed in nonexpressed cells (well exemplified by the odorant genes, which are highly tissue-specific but replicate asynchronously even in fibroblasts ) is also in accord with our results.
Yet, the unscheduled replication of a single locus at the pericentromeric array may interfere with the overall chromosomal structure required for proper chromatid separation and/or segregation. Whatever the mechanism, the two chromatid counterparts in the patients' cells probably separate earlier in mitosis than scheduled, as judged from the far higher levels of monosomy than multisomy. Considering that the mechanisms leading to aneuploidy in the lymphocytes of BCA patients and prostate cancer patients are similar, we assume that the aneuploidy exhibited for chromosome 17 in the BCA cells is non-chromosome-specific and most probably affects other chromosomes as well. We derive this from our observation that the lymphocytes of prostate cancer patients showed increased aneuploidy levels, mostly monosomy of each of five randomly examined chromosomes . This reasoning is supported by findings showing that breast and prostate cancer phenotypes, similar to those of other human cancers, display high levels of aneuploidy, which arise from non-chromosome-specific premature sister chromatid separations .
At the molecular level, loss of one copy of a chromosome is associated with a state of allelic imbalance, often giving rise to the loss of heterozygosity, a phenomenon highly characteristic of the cancerous phenotype. Loss of heterozygosity was first recognized for the RB1 (retinoblastoma) gene and later found in numerous other tumor suppressor genes, including BRCA1 and TP53, both assigned to chromosome 17 (reviewed in Stanbridge ). Whatever the consequences of chromosomal losses in the lymphocytes of patients with solid tumors, these are accompanied by allelic replication timing aberrations, which are non-tumor- and non-chromosome-specific.
Finally, the epigenetic aberrations detected in peripheral blood cells of patients with solid tumors, being associated with the DNA itself rather than its genetically activated products, offer promising biomarkers for cancer detection and monitoring. If these epigenetic changes disappear in breast cancer patients after complete cure from malignancy (after successful surgical and/or chemotherapeutic treatment), women could be provided with further proof and confidence that they are cancer-free.
The authors thank the consenting patients and cancer-free individuals for participating in this study. The authors also thank Jenny Rosensaft for her technical advice.