ROLE OF LIPOXYGENASES IN BREAST CANCER

Rama Natarajan and Jerry Nadler

Department of Diabetes, Endocrinology and Metabolism¹, City of Hope National Medical Center, 1500, E. Duarte Road, Duarte, California 91010, USA.

Received 5/15/98 Accepted 5/29/98

6. EXPRESSION AND REGULATION OF 12-LO IN BREAST CANCER CELLS AND TISSUES.

6.1 Leukocyte-type 12-LO mRNA expression in breast tissue samples.

In order to determine whether 12-LO mRNA expression was altered in breast cancer tissues, we screened six sets of uninvolved and cancer involved breast tissue samples from 6 patients for the presence of the leukocyte-type 12-LO mRNA (46). Total RNA from these samples was subjected to a specific RT-PCR (13,46) to detect 12-LO mRNA levels owing to its the low levels in these tissues. Figures 2A and 2B show Southern blots run with the RT-PCR products from the 12 samples obtained from these 6 patients. Hybridization was performed with a [³²P]-labeled porcine leukocyte 12-LO oligonucleotide probe (upper panels of both Figures 2A and 2B). The size of the expected PCR product was 333 bp. It is clearly seen that in each patient, the cancerous section had a much higher level of 12-LO mRNA expression than the corresponding normal section. In fact, in patients 4-6, 12-LO mRNA was barely expressed in the normal sections. After correction for amplification of the internal control, GAPDH mRNA (PCR product 284 bp), densitometric analysis revealed 3- to 30-fold greater 12-LO mRNA expression in the cancer sections than in the corresponding uninvolved section from the same patients. These results suggest that malignant breast tissues express a much higher level of the 12-LO mRNA in vivo when compared to matched uninvolved tissues.

Figure 2. A) Leukocyte-type 12-LO mRNA expression in matched normal (N) and cancerous (C) breast tissue samples from patients 1-3. B) Same from patients 4-6. Results shown are Southern Blots of the RT-PCR amplified products from the RNA extracted from tissue samples. Hybridization with a leukocyte 12-LO oligonucleotide probe is seen in the upper panels (333 bp) while hybridization with a GAPDH probe to control for PCR amplification efficiency in the corresponding samples is shown in the lower panels (284 bp). The positive control, leukocyte 12-LO cDNA amplification, is seen at the far right. In the negative controls, PCR was run with no RNA. Reprinted with permission from the The Endocrine Society, 1997. J. Clin. Endocrinol. Metabol. 82(6)1790-1798.

In addition to the expected 333 bp 12-LO PCR product, an additional lower band (approximately 300 bp) was observed in all the samples. The identity of this transcript was not clear but since this band was regulated in exactly the same manner as the 333bp band, it may be related to 12-LO mRNA.

6.2 Human 15-LO mRNA expression in breast tissue samples.

Since human 15-LO and leukocyte-type 12-LO are very homologous, we examined the expression of 15-LO mRNA in the same patient tissue samples as above. We used a specific RT-PCR approach which distinguishes between 15-LO and leukocyte-type 12-LO (12,13). Figure 3 is a Southern blot of RT-PCR amplified products to examine 15-LO mRNA expression in normal and cancer tissue sections from the same patients 1, 2, 4, 5 and 6. The positive control, 15-LO cDNA is seen on the far right. The size of the 15-LO PCR product is 333 bp. Strong expression of 15-LO mRNA in human breast tissue and cancer was observed. However, its expression was enhanced in the cancer- involved sections in only two of the 5 samples.

Figure 3. 15-LO mRNA expression in matched normal (N) and cancerous (C) breast tissue samples from patients 1,2,4,5 and 6. Hybridization with a 15-LO oligonucleotide probe is seen in the upper panel (333 bp) while hybridization with a GAPDH oligonucleotide probe (control for PCR) is seen in the lower panel (284 bp). The positive control (15-LO cDNA amplification) is seen on the far right. Reprinted with permission from the Endocrine Society 1997, J. Clin. Endocrinol. Metabol. 82(6)1790-1798.

6.3 12-LO mRNA and protein expression in breast cancer and normal breast cell lines.

Since tissue samples contain a variety of cell types, we also compared leukocyte-type 12-LO mRNA expression in normal versus cancer cell lines. Figure 4A shows a Southern blot of the RT-PCR amplified products from total RNA from two breast cancer cell lines MCF-7 and COH-BR1 as well as an immortal, non-tumorigenic breast epithelial cell line MCF-10F. There was very little basal expression of 12-LO mRNA (333 bp PCR product) in MCF-10F cells (figure 4A). However, distinct expression of 12-LO was seen in the two cancer cell lines MCF-7 and COH-BR1 (7- and 11-fold greater than the MCF-10F cells, corrected for GAPDH mRNA amplification (284 bp) seen in the lower panel). The positive control for PCR, 12-LO cDNA amplification, is seen in the far right. Thus, breast cancer cell lines such as MCF-7 and COH-Br1 had a much higher level of 12-LO mRNA expression relative to the control cell line MCF-10F.

Figure 4. Leukocyte-type 12-LO mRNA and protein expression in breast cancer cell lines and normal breast cell lines. Figure 4A is a Southern blot of the RT-PCR amplified products obtained from 1 microgram each total RNA from two breast cancer cell lines (MCF-7 and COH-BR1) as well as from an immortal non-tumorigenic breast epithelial cell line (MCF-10F). Hybridization was with a leukocyte 12-LO oligonucleotide probe which resulted in the expected 333 bp amplified product. The positive control for PCR, 12-LO cDNA amplification is seen at the far right. GAPDH mRNA amplification (284 bp) is seen in the lower panel. Figure 4B shows an immunoblot to detect leukocyte 12-LO protein in the breast cancer cell lines, MCF-7, MDA-MB-231, T47D and COH-BR1 (lanes 2-5), human vascular smooth muscle cells (lane 6) as well as the primary normal human breast epithelial cell AC113 (alone or treated for 24 hr with EGF 50 ng/ml, lanes 7 and 8). Equal amounts of cell lysates (50 microgram protein) were loaded in each lane. Blots were probed with a peptide antibody to porcine leukocyte 12-LO. Authentic porcine leukocyte 12-LO is seen in lane 1.Reprinted with permission from the Endocrine Society 1997, J. Clin. Endocrinol. Metabol. 82(6)1790-1798.

In addition, we compared 12-LO protein expression in four breast cancer cell lines, MCF-7, MDA-MB-231, COH-BR1 (46) and T47D with that in a normal breast epithelial cell line, AC113, specimen 161 (from Dr. Martha R. Stampfer, University of California, Berkeley) (46). The results of the immunoblot, using a peptide antibody to porcine leukocyte 12-LO (13,24) is seen in Figure 4B. All the 4 breast cancer cell lines showed a band around 75 kD similar to the band seen with human vascular smooth muscle cells on lane 5. Authentic porcine leukocyte 12-LO is in lane 1. The human 12-LO appeared to migrate at a slightly higher molecular size than porcine 12-LO. In contrast, the normal breast epithelial cell line did not show this 12-LO band with or without 24 hr EGF treatment (lanes 6 and 7).

6.4 The effect of EGF on cell-associated 12(S)-HETE levels.

In order to evaluate whether a potent breast epithelial cell growth factor can affect the LO pathway, we examined whether EGF can increase the formation of the 12-LO product, 12(S)-HETE, in MCF-7 breast cancer cells. A 4 hr treatment of the cells with EGF did not affect the levels of released 12(S)-HETE as measured after extraction by a specific radioimmunoassay (12). In contrast, this treatment with EGF (25-100 ng/ml) led to a dose-dependent increase in the levels of cell-associated 12(S)-HETE as seen in Figure 5. EGF also increased intracellular LO enzyme activity (46).

Figure 5. The effect of EGF on the levels of cell-associated immunoreactive 12-HETE in MCF-7 breast cancer cells. Serum-starved MCF-7 cells were treated for 4 hr with EGF in medium containing 0.2% BSA. 12-HETE levels in the cell pellets were quantitated by radioimmunoassay after deacylation and extraction as described (12,24). *, p < 0.01 vs control obtained by analysis of variance (ANOVA).Reprinted with permission from the Endocrine Society 1997, J. Clin. Endocrinol. Metabol. 82(6)1790-1798.

6.5 The effect of EGF on leukocyte-type 12-LO protein expression.

Figure 6 shows that a 36 hr treatment with EGF also led to a marked increase (2-3 fold) in levels of the 12-LO protein (approximately 75 kd). Authentic porcine leukocyte 12-LO is shown in the far left lane. It is noted that the 12-LO in these human MCF-7 cells appeared at a slightly higher molecular weight than the porcine 12-LO.

No clear evidence for the presence of platelet 12-LO in the MCF-7 cells was obtained in these studies either in the basal state, or after treatment with EGF (46). However, the possibility that a platelet-type of 12-LO also contributes to 12(S)-HETE formation in breast tissue samples and in the local environment of tumor cells cannot be completely ruled out.

Figure 6. The effect of EGF on leukocyte-type 12-LO protein expression in MCF-7 cells. Serum-starved cells were treated for 36 hr with EGF. Equal amounts of protein lysates were electorphoresed and subjected to immunoblotting with an antibody raised against to a peptide derived from the sequence of the porcine leukocyte 12-LO. Authentic porcine leukocyte 12-LO protein (Oxford Biomedical Research Co., Oxford, MI) was loaded in the lane on the extreme left. Densitometric representation of the blot is seen in the bar graph below in arbitrary optical density units. EGF (50 and 25 ng/ml) led to significant increases in 12-LO levels (1.8 ± 0.2 fold and 2.3 ± 0.4 fold respectively, both p < 0.001). Reprinted with permission from the Endocrine Society 1997, J. Clin. Endocrinol. Metabol. 82(6)1790-1798.

6.6. The effect of LO and cyclooxygenase inhibitors on the growth of MCF-7 cells.

In order to evaluate the potential functional significance of altered 12-LO expression in the breast cancer cells, we examined the effect of two specific structurally dissimilar 12-LO inhibitors, cinnamyl-3,4-dihydroxy-cyanocinnamate (CDC) and baicalein on the proliferative rates of MCF-7 cells (46). Figure 7 shows that both LO inhibitors led to a marked inhibition of the serum-induced growth of these cells. Since both CDC and baicalein may also block the 5-LO pathway, we also checked the effect of a highly specific 5-LO inhibitor, AA-861. Figure 7 shows that, while this 5-LO inhibitor also had inhibitory effects on the proliferation of the MCF-7 cells, it was not as potent as CDC or baicalein. A cyclooxygenase inhibitor, ibuprofen, at the same concentration as CDC and baicalein, had no significant effect on cell proliferation (figure 7). These results indicate that the LO pathway may mediate, at least in part, the growth of breast cancer cells. However, 12-HETE may not be the only LO product involved in breast cancer and other LO products generated by the 5-LO or other pathways may also play a role.

Figure 7. Effect of lipoxygenase and cyclooxygenase inhibitors on the growth of MCF-7 cells. The inhibitors (10 micromolar each) were added every 48 hr to MCF-7 cells growing in DME medium containing 5% FCS and cell counts obtained on a Coulter counter. Reprinted with permission from the Endocrine Society 1997, J. Clin. Endocrinol. Metabol. 82(6)1790-1798.