Hexavalent Chromium–Induced Chromosome Instability Drives Permanent and Heritable Numerical and Structural Changes and a DNA Repair–Deficient Phenotype

Hexavalent chromium [Cr(VI)] is a known respiratory carcinogen. The most carcinogenic forms of Cr(VI) are the particulates, such as lead chromate, which deposit and persist at lung bifurcation sites in the respiratory tract. These deposited particles dissolve slowly over time, resulting in a chronic exposure of lung cells to Cr(VI) and an accumulation of Cr in tissues (1). Cr-induced tumors are typically found at these bifurcations and are associated with higher tissue burdens of Cr (2, 3). Thus, reoccurring exposure to Cr(VI) is key to its toxic effects.

Lung tumors are characterized by genomic instability. Cr-induced tumors are no exception and are characterized by microsatellite instability (MIN) and chromosome instability (CIN). Markers of MIN were detected in 79% of Cr(VI)-induced tumors, whereas only 15% of nonexposed tumors had increases in these markers (4). In addition, MIN increased as worker chromium exposure increased (4) and was correlated with a decrease in hMLH1 expression (5). However, MIN is considered to occur in cells when they are deficient in mismatch repair. Thus, MIN may play a role in the development of tumors but only after mismatch repair deficiency has developed. Most people exposed to Cr(VI) would be expected to have proficient mismatch repair; thus, it is currently unclear if MIN is a driving factor or a consequence of other changes in the genome.

CIN includes both structural and numerical chromosome abnormalities. Studies have shown Cr(VI)-induced lung tumors exhibit CIN. LOH was observed at 6 different loci in 50% to 75% of Cr(VI)-induced lung tumors; however, this outcome was not significantly different from non-Cr(VI) tumors (4). These findings may implicate LOH as a general mechanism for lung carcinogenesis as most lung cancers exhibit significant CIN (6). In addition, studies of chromate workers have shown increased chromosomal aberrations in cultured lymphocytes (7), as well as increases in binucleated cells (8). This outcome is consistent with cell culture studies showing profound and consistent effects on chromosome structure in cultured cells treated with Cr(VI) (9). Numerical abnormalities have not been assessed specifically in Cr(VI)-induced tumors, but multiple studies show Cr(VI) dramatically alters chromosome number in cultured cells treated with Cr(VI) (10–13). No studies have addressed CIN directly in chromium-induced lung tumors, and no analyses of chromosome complement have been done on Cr(VI)-induced lung tumors.

Cell culture studies have shown prolonged exposure to Cr(VI) (i.e., greater than 24 hours) has significant effects on chromosome stability, inducing increases in aneuploidy (10–12), spindle assembly checkpoint (SAC) bypass (14, 12, 15), centrosome amplification (16), and defects in HR repair (17, 18). All of these can severely affect the stability of chromosome structure and number. Defects in HR repair increase CIN in Cr(VI)-treated cells (19) and predispose them to oncogenic transformation (20). However, no studies have directly addressed chromosome translocations as a result of Cr(VI) exposure. Here, we show particulate Cr(VI) induces permanent chromosome translocations, which arise as a result of defective double-strand break (DSB) repair. In addition, we show Cr(VI) induces permanent alterations in chromosome number in which centrosome amplification may play a role.

WHTBF-6 cells are hTERT-expressing human lung fibroblasts. The cells exhibit a normal diploid karyotype, normal growth parameters, and extended lifespan. They have the same genotoxic and cytotoxic response to metals as their parent primary human lung fibroblasts. WTHBF-6 cells were cultured in a 50:50 mix of DMEM/F12 medium plus 15% cosmic calf serum, 1% l-glutamine, and 1% penicillin/streptomycin. Cells were maintained in a 37°C humidified incubator with 5% CO₂. All cells originated in our laboratory and screened monthly for mycoplasma contamination. Dr. Sandra Wise is a certified cytogenetic technologist and routinely authenticated cells through karyotyping.

Lead chromate was used as a particulate compound and administered as a suspension in water; sodium chromate was used as a soluble chromate compound and dissolved in sterile distilled water as previously described (21).

WTHBF-6 cells were exposed to 5 μg/cm² lead chromate for 24 hours in three independent treatments, each separated by about a month. Workplace concentrations are allowed to reach 5 μg/m³ (22), thus given an average respiration rate and an 8 hour workday, then an occupational exposure would be 33 μg Cr(VI) in a 24-hour period. The cell culture dose used equals 17 μg particulate Cr(VI) for a 24-hour period and, thus, is relevant to human exposure. After each treatment, cells were seeded at colony-forming density, cloned, expanded into cell lines, and then retreated (Fig. 1). Cell lines at each stage were evaluated for chromosomal changes, centrosome number, and ability to repair DNA DSBs. There were 91 control clones and 63 treated clones.

Cells were arrested at metaphase using colchicine and harvested using standard methods for chromosomal analysis (23). Metaphases were Giemsa-banded; at least 20 metaphases were karyotyped for each clonal cell line.

Cells were arrested at metaphase using colchicine and harvested using standard methods for chromosomal analysis (23). A total of 100 metaphase cells were counted for each clonal cell line.

Cells were seeded on glass chamber slides coated with Fibronectin Coating Solution-coating matrix. Cells were then washed with a microtubule-stabilizing buffer, fixed with methanol, air dried and then permeabilized with 0.05% Triton X-100, blocked in centrosome-blocking buffer then incubated with α-tubulin antibody, and washed with PBS, followed by incubation with Alexa Fluor 555 secondary antibody. Finally, slides were incubated with anti-alpha tubulin FITC-conjugated antibody, washed and air dried, and then counterstained with DAPI. Centrosome numbers in 100 mitotic cells were analyzed with an Olympus BX51 fluorescent microscope.

Clones were harvested from subconfluent T25 flasks. Cells were collected by trypsinization, given a hypotonic treatment, fixed with 3:1 methanol:acetic acid. Metaphases were dropped onto clean wet slides and stained with Giemsa. Metaphases were assessed for evidence of SAC bypass including centromere spreading, premature centromere separation, and premature anaphase as defined in previous studies (14). A total of 100 cells were analyzed.

Clones were treated with 1 μmol/L sodium chromate for 24 hours or for 24 hours followed by a chemical-free recovery period of 24 hours. After the given time periods, cells were fixed in 4% paraformaldehyde and permeabilized with 0.2% Triton X-100. Cells were then coincubated with anti–g-H2A.X and 53BP1 primary antibodies, and then coincubated with Alexa Fluor 488 goat anti-rabbit and Alexa Fluor 555 goat anti-mouse secondary antibodies. Nuclei were counterstained with Prolong Gold antifade reagent with DAPI. Slides were analyzed on an Olympus BX51 fluorescent microscope. H2A.X, 53BP1, and colocalized foci were analyzed by eye, and 50 cells per treatment condition were counted.

Anchorage independence is considered the most stringent assay for detecting transformation of cultured cells. To determine anchorage-independent cell growth, control and treated clonal cell lines were suspended in 0.35% agar, plated onto a 0.6% agar base layer in a 60-mm dish at a density of 50,000 cells per dish, and grown for 4 to 6 weeks (24). Cultures were examined microscopically 24 hours after plating to confirm single-cell suspension. Colonies were detected by 5% 4-Nitro 545 tetrazolium chloride staining and counted.

Mean, median, variance, and SD were calculated for all groups and subgroups. Mean values were compared using t tests. Because the distributions of most values were right skewed, a normalizing logarithmic transformation was used for statistical testing. For the small subgroups and the nonnormally distributed subgroups, the Mann–Whitney test was used to compare median values of those subgroups. Proportion values of growth in soft agar and repair deficiency were compared using the χ² and/or Fisher exact tests. P values less than 0.05 were regarded as statistically significant. The statistical analyses were all conducted in SAS 9.4 (SAS Institute).

Previous studies have shown particulate chromate can induce CIN with prolonged treatment; however, it is unclear whether these changes are a permanent state or if these cells are unstable and removed from the cell population over time. To address whether particulate Cr(VI) can induce permanent chromosomal abnormalities, we conducted a clonal expansion experiment as illustrated in Fig. 1. Briefly, cells were treated with lead chromate for 24 hours, allowed to form colonies, which were isolated and expanded into cell lines. Once the first-generation clonal cell lines were established, they were characterized and treated for an additional 24 hours and then again allowed to form colonies and expanded into cell lines forming a second generation of clones. These clones were also characterized and treated for an additional 24 hours to create a third generation of clones for analysis and characterization. Untreated control cells were also expanded and characterized in each generation to ensure the cloning process did not induce changes.

Particulate Cr(VI) induced structural and numerical CIN in treated clones compared with untreated clones (Fig. 2A). This impact was observed at both the whole cell level with 4-fold increase in the percentage of metaphases with CIN for each treated clone, and at the chromosomal level with almost a 6-fold increase in the total number of chromosomes affected (Fig. 2A). There was no increase in either structural or numerical CIN in the control clones above background (parental cells have a background rate of approximately 10% aneuploidy). Lead chromate induced structural alterations in 34% all metaphases across all treated clones. There were a total of 51 structural abnormalities per 100 metaphases, indicating many metaphases had more than one structurally altered chromosome (Fig. 2B). Lead chromate induced numerical CIN or aneuploidy in 43% of all metaphases across all treated clones (Fig. 2C). The most striking numerical change observed was an increase in tetraploidy (defined as 92 ± 4 chromosomes; Fig. 2D). We also considered aneuploidy based on 100 metaphases counted for each clone (Fig. 2E). The average aneuploidy in control clones was 19% compared with 48% in the treated clones. Considered another way, 98% of the treated clones were greater than the average of the control clones (Fig. 2E).

Both the numerical and structural CIN were heritable at the cellular level (Fig. 2A). The first generation had only a slight increase in overall CIN across treated clones. Twenty percent of metaphases had some form of CIN, and there were a total of 21 abnormal events per 100 metaphases. CIN increased in the second generation to 55% of metaphases and a total of 88 abnormal events per 100 metaphases, indicating cells were accumulating multiple structural and/or numerical abnormalities. The third generation had 53% of metaphases in treated cells with some form of CIN and a total of 67 abnormal events in 100 metaphases. There was no increase in CIN across the untreated control generations.

Considering structural and numerical CIN separately, 6% of metaphases from first-generation treated cells had structural abnormalities, which increased to 36% in the second generation and 39% in the third generation. Total structural abnormalities also increased in treated cells with 6, 69, and 52 total alterations per 100 metaphases examined in the first, second, and third generations, respectively (Fig. 2B). There was no increase in control cells.

For numerical CIN, lead chromate induced persistent aneuploidy in all generations with 16%, 46%, and 47% of aneuploid metaphases in the first, second, and third generations, respectively (Fig. 2C). Of these aneuploid cells, 2% of treated clones were tetraploid or near tetraploid; in the second and third generations, 22% and 20% of treated clones were tetraploid or near tetraploid, respectively; whereas less than 1% of control clones exhibited tetraploidy (Fig. 2D). Thus, we saw increases in CIN across all generations in the treated cells but no increase in control cells.

Additional detail regarding specific chromosome changes is seen in Fig. 3. Control clones showed no significant structural changes. Chromosomes 1 (12%), 6 (6%), 7 (5%), 9 (15%), 10 (10%), 13 (15%), and 21 (5%) were the most commonly altered chromosomes in the treated clones (Fig. 3A). The most common chromosome loss was the Y chromosome; 9% of all metaphases analyzed in the treated clones exhibited loss of chromosome Y, compared with 2% in the controls (Fig. 3B). Other chromosomes commonly lost in the treated clones included chromosomes 7 (5%), 17 (5%), 21 (6%), and 22 (7%); chromosome loss in control clones was 3% or less for all chromosomes (Fig. 3B). Chromosome gain (other than ploidy changes) was much less common; the most common chromosome gain for the treated clones was chromosome 13 (Fig. 3C).

Although the analyses above show the overall and cumulative effects of lead chromate on clonally expanded and treated cells, it does not take into consideration the fact that each generation of clones is related to each other. To demonstrate the relatedness of the clones, we assembled a pedigree for each family to show how they are related and how they evolve over time and with additional treatment. We used standard cytogenetic nomenclature according to the International System for Cytogenetic Nomenclature (25); therefore, cells with a single cell missing a chromosome are not reflected here. In some instances, when a clone was treated, the cells had an increased sensitivity to lead chromate and no cells survived treatment; in other instances, only one or two clones survived treatment and successfully expanded into cell lines.

Most control clones had normal karyotypes (Fig. 4A). Clone C1–1 had a balanced translocation between chromosomes 13 and 17 in half of the analyzed metaphases not seen in subsequent generations. Clone C51–1 had 12 normal cells and 8 cells with additional material on chromosome 21. The remaining 89 clones were normal.

Within the treated pedigrees, we considered clones at three levels of analysis: Individual clones, generation, and family. Figure 4B shows the pedigrees of the treated clones. At the level of individual clones, 44% of the clones were abnormal. Specifically, 35 clones had normal karyotypes (T1, T2, T2-2, T21-1, T22-1, T22-3, T3, T3-1, T3-2, T31-2, T31-3, T4, T5, T5-1, T51-2, T51-3, T6, T6-1, T6-2, T6-3, T61-2, T61-3, T62-1, T62-2, T62-3, T63-2, T7, T7-2, T71-2, T71-3, T72-1, T72-2, T72-3, T73-1, and T73-3), and 28 clones had CIN (T1-1, T1-2, T2-1, T2-3, T21-2, T21-3, T22-2, T23-1, T23-2, T23-3, T3-3, T31-1, T32-1, T33-1, T33-2, T33-3, T4-1, T4-2, T41-1, T41-2, T41-3, T51-1, T61-1, T63-1, T7-1, T7-3,T71-1, and T73-2).

Based on generation analysis, all first-generation clones had normal karyotypes, 53% of clones in the second generation were abnormal, and 49% of clones in the third generation were abnormal. Clones that were abnormal in the second generation were also abnormal in the third generation, demonstrating the heritability and persistence of the damage. The specific abnormalities observed in the second generation persisted into the third generation. For example, clone T2-3 had 20 cells with an unbalanced translocation between chromosomes 10 and 13, and 17 of those cells had an extra copy of chromosome 13; in the daughter cells, the same abnormalities persisted but with differing frequencies (T23-1 had 20 cells with the translocation and 3 cells with the extra 13; T23-2 had 20 cells with the translocation and 11 cells with the extra 13; and T23-3 had 20 cells with the translocation and 3 cells with an extra 13). This pattern is also seen between clones T3-3 and T33-1 with both lines being tetraploid, and between T2-1 and two of its daughters T21-2 and T21-3.

In some instances, these third-generation clones acquired additional chromosome abnormalities. For example, clone T4-1 had 91 chromosomes with an unbalanced translocation between chromosomes 1 and 9, and it had one daughter with the same translocation as well as some cells with an isochromosome 7 and another unbalanced translocation between chromosomes 13 and 14. This pattern was also seen between clone T3-3 and two of its daughters (T33-2 and T33-3) and between clone T7-3 and one of its daughters (T73-2).

Analysis of the pedigrees by family showed CIN in all clonal families. In addition, the family pedigrees show five clonal families had clones that were more sensitive to additional Cr(VI) treatment and did not survive to form further generations. In clonal families T1, T4, and T5, there was increased sensitivity in the second generation; specifically, there were only two second-generation clones developed in the T1 family, two second-generation clones in the T4 family, and only one-second generation clone in the T5 family.

Another pattern that becomes apparent with the pedigrees is the induction of a tetraploid chromosome complement. There are four clonal families showing a tetraploid or near-tetraploid state; specifically, families T2, T3, T4, and T7. In two of these tetraploid clones, there is increased CIN in the subsequent generation. For example in the clonal family T3, tetraploidy is the only abnormality observed in one of the second-generation clones. The next generation of this clone results in one clone with the same karyotype; one clone with a tetraploid karyotype and a structural abnormality with chromosome 7; and a clone with a near-tetraploid karyotype with additional loss of either chromosome 5 or chromosome 7 or both. This suggests a tetraploid state may be an initiating event in Cr(VI)-induced CIN. This conclusion is also supported by other clones with a near-tetraploid chromosome complement that have one of the four chromosomes altered as seen in clones T4-1 with daughter clone T41-1 (Fig. 4C) and clones T7-1 with daughter clone T71-1. If the structural alteration had occurred prior to the induction of tetraploidy, then two of the four chromosomes would have the same affected chromosome; we did not observe this in any of the clones.

We also observed clones with a heterogeneous population (not all 20 cells had the same karyotype). Six clonal families have at least one clone that is heterogeneous: T1, T2, T3, T4, T6, and T7. The only family without heterogeneous clones was T5, which consisted of only 5 clones; one was abnormal. In addition, there were two clones (T1-1 and T4-2) with severe CIN in which every cell examined had a different chromosome complement. Clone T4-2 had several cells with the same structural aberrations. Clone T1-1 was highly unstable, with every cell analyzed having different chromosome alterations (Fig. 4C).

Having determined lead chromate–induced permanent CIN, we wanted to determine if this state was sufficient for neoplastic transformation. Normal human lung cells exhibit anchorage-dependent growth and require a surface that allows them to attach to, flatten out, and divide. Transformed cells do not require a surface to attach to and have the ability to grow in suspension. Anchorage-independent growth strongly correlates with tumorigenicity and invasiveness in several cell types, including small-cell lung carcinomas (20). Therefore, we assessed the ability of the clones to grow in soft agar. Figure 5A shows pictures of colonies in agar and microscopically. Eighteen percent of all treated clones developed the ability to grow in soft agar (Fig. 5B). Considered by generation, 14.4% of first-generation clones grew in agar, 12.5% of second-generation clones grew in agar, and 20.5% of third-generation clones grew in soft agar (Fig. 5C). None of the control clones grew in agar.

Comparing the karyotype results with the agar results raises interesting questions about the frequency with which cells with CIN grow in soft agar and the frequency with which cells that grow in soft agar exhibit abnormal karyotypes. First, considering the clones that grew in agar, 50% were chromosomally abnormal; none in the first generation, 100% in the second generation. All of the clones in the third generation exhibited some degree of CIN, but only 50% had consistent CIN reflected by the same chromosomes being affected in at least 85% of the cells analyzed (Table 1). Second, considering the percentage of chromosomally abnormal clones that grew in agar, 14% of the chromosomally abnormal clones grew in agar; none in the first generation, 25% in the second generation, and 15% in the third generation (Table 1). Thus, CIN is a common phenotype in clones neoplastically transformed by particulate Cr(VI), but several with particulate Cr(VI)-induced CIN do not grow in agar.

To begin to understand the underlying causes of the observed CIN, we investigated potential contributing factors to both numerical and structural CIN. We considered centrosome amplification. Abnormal centrosome numbers are common in tumors (26, 27). Incorrect centrosome number can lead to multipolar segregation of chromosomes resulting in the incorrect chromosome complement in daughter cells (28). In addition, multipolar cells can abort cytokinesis and result in a single tetraploid cell (29, 30). Twenty-nine percent of all treated clones exhibited centrosome amplification (Fig. 6A). Considered by generation, 43% of first-generation clones had amplified centrosomes, 44% of second-generation clones had amplified centrosomes, and 21% of third-generation clones had amplified centrosomes (Fig. 6B). None of the control clones showed amplified centrosomes. Broken down by generation, we saw 3 of 7 clones in the first generation, 7 of 16 clones in the second generation, and 21 of 39 clones in the third generation exhibited centrosome amplification.

Another aspect that may contribute to numerical CIN is defects in the SAC. The SAC serves to ensure that cells do not progress to anaphase and cytokinesis until all of the chromosomes are properly aligned. Thus, we analyzed metaphases for indication of SAC bypass manifested as centromere spreading, premature centromere division, and premature anaphase. Forty-one percent of all treated clones exhibited SAC bypass (Fig. 6C). Considered by generation, 29% of first-generation clones exhibited SAC bypass, 44% of second-generation clones exhibited SAC bypass, and 42% of third-generation clones exhibited SAC bypass (Fig. 6D).

In order to better understand the structural alterations observed, we sought to reveal the underlying contributing factors leading to chromosome translocations. Chromosome translocations require multiple DNA DSBs and thus we chose to look at the efficiency of DNA DSB repair. Previous studies from our lab show Cr(VI) induces persistent DNA DSBs (17). We treated cells with soluble sodium chromate for 24 hours and measured the induction of DNA DSBs using colocalized g-H2AX and 53BP1 foci. In complimentary dishes, we washed the treatment out and gave cells 24 hours to repair the damage. Figure 6E shows a representative example of repair deficiency in a control clone and a treated clone; after 24 hours of treatment, the control and treated clones show similar levels (22 and 24, respectively) in the percentage of cells with more than 5 foci, indicating DNA DSBs occurred. When the cells were given time to repair, we found the control clone showed a reduced number of foci (10%) indicating repair, whereas the treated clone failed to repair the DNA DSBs and actually had more damage (40%). Figure 6F shows 59% of all treated clones displayed some level of repair deficiency. Considered by generation, 29% of treated clones were repair deficient in the first generation, 63% in the second generation, and 64% in the third generation (Fig. 6G). None of the control clones were repair deficient.

CIN is thought to be a central component of the carcinogenic mechanism for Cr(VI) as the model of clonal expansion of driver mutations does not fit well with the available data. Most studies of the carcinogenic mechanism of Cr(VI) involve data that consider impacts in the immediate aftermath of exposure, i.e., assays are treated with Cr(VI) and the outcomes measured immediately after exposure. Although this approach indicates impacts of exposure, it cannot inform whether those impacts are transient changes that require the presence of Cr(VI) or if they are permanent changes that persist long after the exposure has ended. The notable exception being assays for transformation, which occur weeks after exposure ends. The cell lines we analyzed were all clonal originating from a single cell. Thus, the observed outcomes represent permanent changes. The second generation of cells was derived from the first, and the third from the second allowing for us to see changes heritable at a cellular level. Our data show all of our endpoints, CIN, DNA repair deficiency, centrosome amplification, and growth in soft agar were permanent and heritable changes.

This article is the first to report Cr(VI)-induced translocations. It is well established that Cr(VI) induces chromosomal aberrations including chromatid and chromosome lesions. Observations of dicentrics, triradial figures, and chromatid exchanges indicate Cr(VI) has the potential to induce translocations (23, 31, 32), but the only report of any chromosome specificity is limited to observations that aberrations preferentially occur in euchromatic regions over heterochromatic regions (33). We found the amount of structural chromosome damage was heritable and increased with generation, indicating progressively more CIN. This outcome is consistent with previous observations of increasing chromosomal aberrations with increasing length of Cr(VI) exposure in solid stained chromosomes from human lung cells (10–12) and further supports the hypothesis that CIN is a significant part of Cr(VI)'s carcinogenic mechanism.

Chromosome translocations, deletions, and duplications are the most frequent structural rearrangements observed in somatic tumors, but distinct and recurrent chromosome translocations in tumors are rare (34). We found Cr(VI) caused recurrent translocations in chromosome numbers 1, 3, 6, 7, 9, 10, 13, 17, and 21. Moreover, of the recurring chromosomal translocations observed, 94% were nonreciprocal or unbalanced translocations leading to either the loss or gain of partial chromosomes. Interestingly, all but one of the recurring translocations observed were whole-arm translocations resulting in loss of chromosomes 1p, 3p, 9p, 10p, 16q, and 17p or gain of chromosomes 9q and 13q. There are few reports of recurring whole-arm translocations in any cancer. Whole-arm translocations are frequent in head and neck squamous cell carcinoma with whole-arm losses seen at 3p, 8p, 9p, 17p, and 18q and gains at 3q, 5p, 7p, 8q, and 20q (35). Whole-arm translocations leading to loss of 17p have been implicated in advanced hematologic malignancies with poor prognoses (36). Whole-arm translocations involving chromosomes 7, 10, 12, 14, and 21 were report for cervical cancers (37). One study investigated whole-arm chromosome translocations among adenocarcinoma versus squamous cell carcinoma and found the whole-arm translocations observed in squamous cell carcinomas were derived from centromere breakage and refusion (38). Chromium causes predominately squamous cell carcinoma in lungs of exposed workers. Thus, Cr(VI) may induce centromeric instability, which is also supported by our previous observations of premature centromere division in Cr(VI)-exposed cells (12, 14).

DNA DSBs have emerged as a critical step in the carcinogenic mechanism for Cr(VI) and underlies the structural chromosomal changes (20). Remarkably, data show, in addition to inducing these breaks, Cr(VI) inhibits their repair, specifically targeting the RAD51 step of homologous recombination (HR) repair (17, 18, 39). Loss of HR allows Cr(VI) to induce the neoplastic transformation of cells and increase levels of CIN. In this report, we find for the first time that Cr(VI) inhibition of DNA repair persists long after exposure has ceased and is heritable at a cellular level. This outcome further strengthens the importance of DSB repair inhibition in the carcinogenic mechanism of Cr(VI).

We found Cr(VI)-induced aneuploidy, which is a consistent hallmark of cancer cells (40). We found Cr(VI) caused amplification or loss of chromosome numbers 7, 13, 17, 21, 22, and the Y chromosome, all have been observed in lung cancer (41). Loss of chromosome Y has been associated with increased risk of nonhematologic cancer in men (42). Loss of chromosome 7 has been suggested as a marker for melanoma due to the loss of EGFR located at chromosome 7p12.3-p12.1 (43). Monosomy of chromosome 17 has been implicated in breast cancer with loss of p53, BRCA1, and TOP2A (44). LOH of chromosome 22 has been associated with multiple cancers, including lung cancer (45), and suggests a loss of tumor-suppressor genes (46). Interestingly, we found a higher proportion of small chromosomes were affected compared with larger chromosomes, which is consistent with a study considering 43,205 human tumors with greater loss of small chromosomes (47). Gain of chromosome 13 has been implicated in colorectal cancer (40). These outcomes are also consistent with previous observations showing Cr(VI) induces aneuploidy in human lung cells including hypodiploidy, hyperdiploidy, and tetraploidy (9, 12, 14, 23) and further supports the hypothesis that CIN is a significant part of Cr(VI)'s carcinogenic mechanism.

We found an increase in tetraploidy and near tetraploidy. This outcome is consistent with observations in lung cancers with severe aneuploidy (6). Several recent studies have shown tetraploidy is often found in precancerous lesions in a variety of tissue types (40, 48). Studies have also shown tetraploidy as an intermediate step in chemically induced aneuploidy and cellular transformation (24, 49, 50).

Centrosome amplification and SAC bypass have emerged as a factor in the carcinogenic mechanism for Cr(VI) and underlie the numerical chromosomal changes (11, 12, 14). Cr(VI) induces centrosome amplification in human lung cells, which is maintained in Cr(VI)-transformed cells (24). The target appears to be the centrioles within the centrosomes with alterations of centriole numbers and an increase in centriole splitting (16). In this report, we find for the first time Cr(VI)-induced amplification persists long after exposure has ceased and is heritable at a cellular level. In addition, we found SAC bypass persisted and was heritable. These outcomes further strengthen the importance of centrosome amplification and SAC bypass in the carcinogenic mechanism of Cr(VI).

Human pathology studies of workers with Cr(VI)-induced lung cancer show Cr accumulates and tumors form at bronchial bifurcation sites where Cr(VI) particles affect, persist, and dissolve. This accumulation of Cr was noted as a key factor in the incidence of tumors in these workers, more so than the dose. Our data are consistent with this conclusion as they show the majority of effects do not fully manifest themselves until after two exposures to Cr(VI). Our previous data would also appear to implicate accumulation as the key factor in those studies and the major impacts were not seen until after at least 48 hours of exposures.

In sum, we show for the first time Cr(VI) induces chromosome translocations, and the chromosome translocations, aneuploidy, and polyploidy observed are permanent and heritable. Further, we show the underlying centrosome amplification and DNA damage repair defects are also permanent and heritable. These chromosome imbalances likely lead to preferential selection and survival of cells in which oncogenes are activated and/or tumor-suppressor genes are lost providing a growth advantage for cancerous cells.

No potential conflicts of interest were disclosed.

Conception and design: S.S. Wise, J.P. Wise, Sr.

Development of methodology: S.S. Wise, J.P. Wise, Sr.

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): S.S. Wise, J. Martino, J.P. Wise, Sr.

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S.S. Wise, A.E.-M. Aboueissa, J. Martino, J.P. Wise, Sr.

Writing, review, and/or revision of the manuscript: S.S. Wise, A.E.-M. Aboueissa, J. Martino, J.P. Wise, Sr.

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S.S. Wise, J.P. Wise, Sr.

Study supervision: S.S. Wise, J.P. Wise, Sr.

This work was supported by a grant from the National Institute of Environmental Health Sciences (ES016893 to J.P. Wise, Sr.). The authors would like to thank Rachel Speer, Greer Chapman, Kelsey Thompson, Therry The, Hong Xie, Christy Gianios, Jr., Kelly Holland, Aaron Howell, and Blair Cade for technical and administrative assistance.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.