Aberrant CBFA2T3B gene promoter methylation in breast tumors

Additional File 1:

CBFA2T3 gene expression levels assayed using real-time RT-PCR The raw data CBFA2T3B expression levels and preliminary CBFA2T3A expression levels in breast tumor cell lines are shown. The y-axis represents the fluorescence detection scale and the x-axis represents the C_T of amplification. The CBFA2T3 gene expresses at endogenously low levels and requires at least 500 ng of reverse transcribed total RNA to cDNA template for reproducible detection. In contrast, the housekeeping genes such as CYPA require only 100 ng of template. When 500 ng is used, the CYPA expression levels are off the fluorescence scale. Note the C_T values are below 35 cycles for the down-regulated cell lines such as MDA-MB-231. This low-level of expression is undetectable by conventional RT-PCR. Moreover, because of this low endogenous expression the CBFA2T3 mRNA could not be reliably detected using Northern Blots or RNase protection (pdf file).

Format: PDF Size: 223KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 2:

CBFA2T3B promoter methylation levels examined using MSP The methylation levels in 24 breast tumor cell lines (labeled), 20 primary breast tumors (1T-20T), 20 normal breast counterparts (1N-20N) and 24 normal whole blood samples (1–24) are shown. Four separate regions (1–4) spanning 1-kb of CBFA2T3B promoter sequence were amplified using primers to detect for either unmethylated or methylated cytosines (see Figure 1B for primer locations and Additional file 7 for primer sequences and annealing temperatures). The asterisks indicate the samples examined by bisulfite sequencing. Unmethylated (U) and methylated (M) band intensities were scored as high (black dot), low (gray dot) or negative (white dot). Low-level basal methylation defined by high unmethylated to low methylated band intensities was detected in all normal bloods in ≥ 2/4 regions. High methylation was also detected in 21% of bloods in region four. Samples were considered hypermethylated when high intensity bands amplified in all four regions or hypomethylated when no methylation was detected. Based on this, 21% of breast tumor cell lines were hypermethylated (e.g. MDA-MB-231 and MDA-MB-468), 16% hypomethylated (e.g. BT-483 and MDA-MB-361) and 63% displayed a combination of basal to high methylation. Primary breast tumor samples were complex. 57% displayed basal methylation in ≥ 2/4 regions, 14% displayed high methylation in ≥ 2/4 regions and 29% tested negative in ≥ 3/4 regions. Normal breast counterpart samples were similar with 62% displaying basal methylation in ≥ 2/4 regions, 14% displaying high methylation in ≥ 2/4 regions and 24% testing negative in ≥ 3/4 regions. The detection of high-methylation in MCF-7 is due to spurious amplification events. Several overlapping primers spanning this region in this cell line were methylation negative (data not shown). This cell line is hypomethylated according to real-time MSP (pdf file).

Format: PDF Size: 435KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 3:

CBFA2T3B promoter methylation levels assayed using second-round real-time MSP CBFA2T3B promoter methylation levels were assayed using second-round real-time MSP. The raw data methylation levels in normal whole blood samples and breast tumor cell lines are shown. The y-axis represents the fluorescence detection scale and the x-axis represents the C_T of amplification. Second round real-time MSP was performed on a bisulfite sequencing amplicon using internal forward primers to detect for either unmethylated (uF) or methylated (mF) cytosines at the Sp1 CpG sites shown in Figure 1C. The C_T for methylated amplification in breast tumor cell lines was highly aberrant compared to normal blood samples. Note the late unmethylated C_T obtained in MDA-MB-231 compared the early hypermethylated C_T. In contrast, BT-483 shows an early unmethylated and late hypomethylated C_T (pdf file).

Format: PDF Size: 340KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 4:

CBFA2T3B promoter methylation melt curve analysis CBFA2T3B promoter methylation melt curves were examined following second-round real-time MSP. Raw data melt curves of second round amplicons in normal blood samples and breast tumor cell lines are shown. Curves were calculated from the negative derivative in fluorescence over temperature versus temperature (-dF/dT_m versus T_m). Normal blood samples displayed consistent peak levels for both the unmethylated and methylated mlcls. In contrast, breast tumor cell lines displayed highly aberrant peak levels depicted by broad melt transitions and heterogeneous melt curves reflective of the aberrant concentration and composition of 5-methylcytosines (pdf file).

Format: PDF Size: 260KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 5:

CBFA2T3B promoter methylation levels versus gene expression Methylation indices were calculated as methylated promoter mlcls per 10⁴ cells for breast tumor cell lines with pre-determined 16q24.3 DNA mlcls per cell and plotted against CBFA2T3 mRNA mlcls per 10⁴ cells. The y-axis represents methylation levels assayed using real-time MSP and the x-axis represents expression levels assayed using real-time RT-PCR. Both data sets are shown on a log scale. Each black circle represents a different breast tumor cell line. The asterisk marks the median methylation and median gene expression levels. The median methylation index of 0.02 (i.e. 2 mlcls or alleles methylated in 100) is calculated from the median methylation level of 450 methylated alleles per 10⁴ cells divided by the number of unmethylated 'active' alleles in the 10⁴ cells or 20,000 alleles, i.e. [450 ÷ (20,000 - 450) = 0.02]. The median gene expression index of 0.2 (i.e. 20 mRNA mlcls expressed in 98 'active' alleles) is calculated from the median expression level of 4,500 mRNA mlcls per 10⁴ cells divided by the number of unmethylated 'active' mlcls in the 10⁴ cells, i.e. (4,500 ÷ 19,550 = 0.2). This calculation equates to approximately 4–5 mRNA mlcls expressed per 10 cells and suggests that the CBFA2T3B gene is largely transcriptionally inert. The remaining active alleles may be trans-factor dependent for expression. An inverse correlation between promoter methylation and expression levels per population of cells was established (r² = 0.77; r = -0.9, P < .0001). In hypermethylated MDA-MB-231, approximately 17,000 promoter mlcls were methylated (i.e. mi = 0.85 as 17,000 in 20,000 are methylated) and 4 mRNA mlcls expressed per 10⁴ cells. In hypomethylated BT-483, approximately 5 promoter mlcls were methylated (mi = 0.0002) and 120,000 (± 40,000) mRNA mlcls expressed per 10⁴ cells. This elevated expression equates that 12 (± 4) mRNA mlcls per cell are expressed from an estimated four-promoter mlcls per cell (i.e. four-16q24.3 DNA mlcls per cell) or that one 'active' unmethylated CBFA2T3B promoter mlcl per cell transcribes 2–4 mRNA mlcls. From a methylation index of approximately 2% and greater, a large increase in y may lead to a small decrease in x. Under this condition, the relationship may be asymptotic. The regression equation y = 25,801 x^-0.58 (r² = 0.7669) describes the methylation and expression relationship. When plotted as methylation index values the equation was y = 1.05 x^-0.61 (r² = 0.7665) (pdf file).

Format: PDF Size: 33KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 6:

CBFA2T3B promoter methylation levels assayed using second-round real-time MSP The methylation levels in normal whole blood samples, normal breast counterparts, primary breast tumors and breast tumor cell lines were assayed at the Sp1 site shown in Figure 1C and plotted as absolute methylation ratios (u/m). All samples are labeled corresponding in part to Additional file 2. The asterisks indicate the samples examined by bisulfite sequencing. The gray highlights indicate the normal blood and normal breast counterpart 'full' methylation ranges. Basal methylation ratios in normal bloods averaged 60:1 (cumulative mean) unmethylated to methylated CBFA2T3B promoter mlcls. This average coincided with conventional MSP band intensities of 100:1 based on pUC19 DNA/MspI marker concentrations (see Figure 3) and the median methylation index of 0.02 (i.e. 2 mlcls in 100 are methylated). Normal blood ratios ranged 20:1 to 160:1 unmethylated to methylated mlcls. Normal breast counterparts were similar to normal bloods averaging 100:1 but with a larger range of 10:1 to 350:1. Relative to the normal samples, breast tumors displayed highly aberrant methylation ratios clearly resolved by second-round real-time MSP. 75% of breast tumor cell lines were aberrantly methylated outside the full range of normal blood basal methylation. 58% were outside the range of normal breast. Half of the aberrations were either hypo or hypermethylated relative to both normal blood and normal breast. Similar to cell lines, 51% of primary breast tumors were aberrantly methylated relative to normal blood. 35% were aberrant relative to normal breast (i.e. 24% were hypermethylated and 11% hypomethylated). Aberrant methylation ratios ranged from 1:10 in hypermethylated MDA-MB-231 to 7,000:1 in hypomethylated BT-483 (pdf file).

Format: PDF Size: 103KB Download file

This file can be viewed with: Adobe Acrobat Reader

Additional File 7:

Primers, probes and annealing temperatures used (pdf file).

Format: PDF Size: 22KB Download file

This file can be viewed with: Adobe Acrobat Reader