Increased Proteasome Activity, Ubiquitin-Conjugating Enzymes, and eEF1A Translation Factor Detected in Breast Cancer Tissue

The ubiquitin (Ub)/proteasome proteolytic pathway is required for efficient cell cycle control, stress response, DNA repair, and differentiation (1–3). Mutations in this pathway can cause pleiotropic defects because of its involvement in virtually all aspects of cell function. Consequently, the characterization of the Ub/proteasome system for the development of treatment for cancer and other malignancies is an area of active investigation (4–7).

The 26S proteasome consists of a catalytic (20S) particle, within which reside three distinct hydrolytic activities. The structure and function of the 20S catalytic particle is conserved in evolution, and its compartmentalized organization ensures that the proteolytic activities are sequestered within the interior of the proteasome (8). A large 19S regulatory particle interacts with the 20S particle to facilitate recognition, unfolding, and degradation of ubiquitinated substrates (9, 10). Ub is covalently attached to lysine side chains in cellular proteins (11). The ligation of Ub to proteins requires the action of three enzymes termed Ub activating (E1), Ub conjugating (E2), and Ub ligases (E3; ref. 1). The sequential addition of Ub moieties results in the formation of a multi-Ub chain, which facilitates protein degradation by promoting interaction of substrates with the proteasome (12, 13).

Malignant conditions are frequently associated with altered abundance and stability of regulatory proteins. Therefore, it is likely that the expression of a unique repertoire of proteins underlies the transition from normal to abnormal growth. Proteasome activity is often elevated in human cancers (14, 15), consistent with the requirement for this proteolytic system in rapidly growing tissues. In an attempt to understand the underlying biochemical function of the Ub/proteasome pathway in cancer, we examined its activity in patient-matched breast cancer and control tissues (obtained from the Cancer Institute of New Jersey Tissue Retrieval Service, New Brunswick, NJ). These studies revealed a striking uniformity in the activation of this proteolytic system in breast cancer cells. Despite considerable variance among individual breast cancer tissue specimens, >90% of the positively diagnosed breast cancer samples displayed convincing evidence for activation of the Ub/proteasome pathway. We also determined that the levels of the UbcH1, UbcH2, and UbcH5 Ub-conjugating enzymes were increased in breast cancer, but not in colon cancer tissues.

It was recently shown that the cotranslational degradation of newly synthesized misfolded proteins requires the Ub/proteasome system (16–18). We determined that the translation elongation factor eEF1A is required for the efficient degradation of nascent polypeptide chains, especially in ATP-depleting conditions, and in the presence of protein synthesis inhibitors (19). Because the frequency of translation errors might increase in actively growing cells, and under conditions of stress, we investigated if the levels of eEF1A were altered in breast cancer. Immunoblotting studies revealed dramatic increase in eEF1A levels, suggesting that defects in protein synthesis quality control are associated with breast cancer. Because eEF1A expression is also increased in other cancers (20, 21), it is likely that this result reflects a more general response to aberrant growth (22), unlike the specific expression of certain Ub-conjugating enzymes in breast cancer. Significantly, none of our findings were reproduced in a well-defined cell culture system that is used as a model system for breast cancer (23, 24). Based on the findings reported here, we propose that the characterization of primary tissue specimens will be required to accurately define the role of the Ub/proteasome pathway in breast cancer.

Primary cancer tissue specimens. All tissue samples (breast and colon cancer) were purchased from the Cancer Institute of New Jersey Tissue Retrieval Service. Specimens were obtained in pairs, which represented previously diagnosed breast/colon cancer and patient-matched normal adjacent tissue. A pathology report accompanied each specimen and provided a characterization of the sample (see Supplemental Table). Two benign tumor samples were also obtained, although only one included a patient-matched control specimen. Benign tumor samples were typically not archived; therefore, only limited samples were available.

Preparation of protein extracts. Each tissue sample was suspended in 2 mL lysis buffer [50 mmol/L HEPES (pH 7.5), 150 mmol/L NaCl, 5 mmol/L Na-EDTA] that contained protease inhibitors (Roche, Indianapolis, IN). Proteasome inhibitors were not added to these tissue extracts. Tissues were kept on ice and disrupted using a tissue homogenizer (Polytron, PT 3100; Brinkmann Instruments, Inc., Westbury, NY) using brief pulses, for a total of 30 to 45 seconds. Lysates were further diluted into 2 mL lysis buffer containing 1% Triton X-100, and then centrifuged at 12,000 × g for 20 minutes at 4°C. Protein concentration was determined using the Bradford reagent (Bio-Rad, Bethesda, MD).

Antibodies and immunoblotting. Antibodies against Rpt1, Rpn2, and 20S core α-subunits, were purchased from Affiniti, Inc. (Exeter, United Kingdom). Antibodies against Ub-conjugating enzymes were from Boston Biochem (Cambridge, MA), and Ub antibody was purchased from Sigma (St. Louis, MO). Antibodies against eEF1A were obtained from Upstate USA (Charlottesville, VA). Equal amounts of protein extract (30 μg) were resolved in SDS-10% polyacrylamide gels (SDS-PAGE), and transferred to nitrocellulose (Hoefer TE70-Semi-Dry; GE Healthcare, Piscataway, NJ). The filters were sequentially incubated with antibodies against Rpt1, UbcH5, and Ub. Due to the high-level expression of eEF1A, we resolved 10 μg of the same protein preparations in a separate polyacrylamide gel. The reactions were detected with enhanced chemiluminescence (NEN/Perkin-Elmer, Boston, MA) and quantified using Kodak 1D densitometry software (Kodak, Rochester, NY).

Proteasome activity determination. Equal amount of lysate was mixed with 1 mL proteasome assay buffer [25 mmol/L HEPES (pH 7.5), 0.5 mmol/L EDTA], containing 40 μmol/L fluorogenic substrate SUC-Leu-Leu-Val-Tyr-AMC (Boston Biochem). The fluorescence measurements represented chymotrypsin-like activity, and were determined using a Turner 7000 flurometer.

Proteasome activity is increased in breast cancer. We obtained a collection of human primary breast cancer specimens (25 individual samples) that were accompanied by patient-matched controls (normal adjacent tissue). Patient age ranged from 35 to 89 years (with a median of 47 years), and a single sample was obtained from a 62-year-old male (sample 3). Tissues were homogenized and proteasome activity was determined using an equal amount of protein extracts. Proteasome activity for each breast cancer sample was adjusted to the activity that was present in patient-matched control specimens (that were set to an arbitrary value of 100). We detected significantly higher proteasome activity in 23 of 25 positively diagnosed samples (P < 0.001; Fig. 1A). Only one specimen (sample 2) displayed reduced activity and a single isolate showed no significant change (sample 8). Because of the broad range in proteasome activity in the various patient tissue samples, we categorized subgroups and determined that only three specimens had <2-fold increased proteasome activity than the control tissue. Ten specimens displayed activity that was 2- to 5-fold higher than the patient-matched control (mean value ∼3-fold increase; P < 0.001), seven specimens had 5- to 10-fold higher levels (mean value ∼8-fold increase; P < 0.001), and five specimens displayed a remarkable 10- to 35-fold elevated proteasome activity (mean value ∼20-fold increase; P < 0.001). It is significant that two additional specimens that represented benign solid tumors showed no increase in proteasome activity. Although both benign solid tumor samples contained very similar proteasome activities, only one specimen (see sample 9) was available with a patient-matched control sample, and therefore the second specimen (sample 10) was not plotted (Fig. 1A). The observation that proteasome activity in benign tumors was indistinguishable from control levels provides compelling support for the hypothesis that the activity of the Ub/proteasome pathway is activated specifically in neoplastic cells. The basal proteasome activity in control tissues varied over an ∼3-fold range (with an average value of 144 ± 101), whereas the activity in 25 breast cancer specimens varied over an ∼13-fold range (with an average value of 699 ± 489). Therefore, absolute proteasome activity is increased ∼5-fold in breast cancer (before correction against the patient-matched control tissue). The variance in activity measurements in cancer tissues might reflect differences in disease stage, cellular origin of the malignancy, and condition of the tissue samples during and following collection. A pathology report that accompanied the tissue specimens provided the gender and patient's age, likely cellular origin, degree of infiltration, and the results immunohistochemistry analysis. The majority of specimens (20 of 25) were derived from invasive ductal carcinomas, although metastatic adenocarcinoma (1), intraductal carcinoma (1), lobular carcinoma (2), and mucinous carcinoma (1) were also examined. The tissues were graded with regard to invasiveness and involvement of lymph nodes. A determination was also provided to indicate estrogen receptor status and Her2 antigenicity. An inspection of these data showed that the highest proteasome activities were observed in infiltrated ductal carcinomas, although the values among these specimens were highly variable, which could reflect the large sample size (n = 20; mean increase = 8-fold; P < 0.001). High activity was also detected in one metastatic adenocarcinoma specimen (16-fold), whereas tissues displaying fibrocystic changes were not accompanied by significant increase in proteasome activity (sample 9, ID-3576; sample 10, ID-3505). Some surprises were evident. For example, sample 16 (ID-3353), which represented an infiltrating ductal carcinoma, displayed the highest proteasome activity (>30-fold) and high levels of UbcH5 (>40-fold) and eEF1A (>40-fold). These cells were poorly differentiated, and 13 of 16 lymph nodes were affected. This specimen was accompanied by low basal activity in control tissue. Yet, immunohistochemical analysis of these cells was entirely negative. Collectively, these data suggest that monitoring the activity of the Ub/proteasome pathway, and the expression of UbcH5 and eEF1A, could provide important insight that could complement conventional diagnostic approaches. However, the value of the pathology and histochemical inspection is evident, because the cellular origin, extent of infiltration, and disease progression cannot be determined solely through biochemical approaches.

Increased proteasome activity is related to high-level expression of proteasome subunits. To determine if the elevated proteasome activity in breast cancer was the result of higher expression of proteasome subunits, in contrast to posttranslational activation (for instance by phosphorylation; ref. 25), we resolved the protein extracts in an SDS-polyacrylamide gel and incubated an immunoblot with antibodies against the proteasome subunit Rpt1, which is located in the 19S regulatory particle (9). Consistent with the findings described above, we detected higher levels of Rpt1 (Fig. 1B) in most samples that had higher proteasome activity (Fig. 1A). One sample was independently examined twice (indicated as 4a and 4b). The same immunoblots were incubated with antibodies against subunits in the 20S catalytic particle, and increased levels were detected in breast cancer specimens (data not shown). The two benign tumor specimens (samples 9 and 10) did not display elevated levels of Rpt1, and, furthermore, we did not observe a single instance where the control sample had higher proteasome levels than the adjacent cancer tissue. However, the relationship between proteasome abundance and activity is not linear, indicating that other factors, such as posttranslational modifications, could also increase proteasome activity. We detected higher Rpt1 levels in 20 of 25 specimens, and only sample 8 displayed higher Rpt1 levels without a corresponding increase in proteasome activity. In contrast, samples 14 and 15 showed ∼2.5-fold increased proteasome activity, without a change in Rpt1 levels, whereas sample 2 showed no increase in either abundance or activity. Further characterization of these samples showed higher eEF1A levels in specimen 15. Sample 8 is a notable exception in the set of 25 breast cancer specimens because it did not display higher proteasome activity or eEF1A levels (Fig. 4), although it expressed higher levels of UbcH5 (Fig. 2B) and multi-Ub proteins (Fig. 2A). Collectively, these findings show that by monitoring both proteasome activity, and abundance of UbcH5 and eEF1A, the majority of breast cancer specimens can successfully be identified.

High molecular weight ubiquitinated proteins accumulate in breast cancer. The majority of substrates that are degraded by the 26S proteasome are ligated to multi-Ub chains to promote their targeting to the proteasome. If proteasome activity were related to the levels of proteolytic substrates, one might detect increased abundance of multiubiquitinated proteins in breast cancer tissues. Therefore, we compared the levels of multiubiquitinated proteins in breast cancer and control tissue extracts. A subset of the results is shown (Fig. 2A). Immunoblots were incubated with antibodies against Ub and, as expected, higher levels were detected in breast cancer specimens 6, 7, 8, and 11. Similar results were observed in the majority of breast cancer samples (data not shown). The levels of ubiquitinated proteins were not increased in specimen 5, consistent with the lower proteasome activity in this tissue specimen (Fig. 1A). Specimens 9 and 10 represent benign solid tumor protein extracts and, in agreement with the results in Fig. 1, there was no evidence for higher levels of ubiquitinated proteins. The increased proteasome activity, and abundance of multi-Ub proteins in breast cancer tissues, probably reflects a cellular response to persistent and high levels of damaged (ubiquitinated) proteins. Because cancer cells are irrevocably changed, the increased proteasome activity and abundance could reflect a compensatory response to the altered cellular condition.

The expression of the UbcH5 ubiquitin-conjugating (E2) enzyme is increased in primary breast cancer tissue. Based on the higher levels of multiubiquitinated proteins in breast cancer tissue extracts, we determined if the levels of an abundant Ub-conjugating (E2) enzyme were altered. The Ubc4 class of E2 enzymes plays an important role in the turnover of damaged proteins, and the expression of this enzyme is rapidly induced in yeast, under conditions of stress (26). Protein damage caused by heat stress and protein synthesis inhibitors results in rapid association between the proteasome and Ubc4 (26). Because of the rapid proliferation of cancer cells, we investigated if the expression of UbcH5 (the human counterpart of Ubc4) was altered in breast cancer. Protein extracts were prepared from breast cancer and control samples, resolved in SDS-PAGE, and an immunoblot was incubated with antibodies against UbcH5. We observed marked increase in the levels of UbcH5 in most breast cancer samples (Fig. 2B). However, UbcH5 was not detected at higher levels in any of the control tissues or benign solid tumors. Many breast cancer samples displayed >8-fold higher levels of UbcH5, and specific specimens showed ∼17-, 22-, 27-, and 44-fold higher levels. Specimen 4 was examined twice (4a and 4b) and similar results were observed, demonstrating that the immunoblotting results were reproduced faithfully. In agreement with the earlier results, the failure to detect higher levels of UbcH5 in benign tumor tissue samples (samples 9 and 10) is consistent with the results in Fig. 1 and supports the hypothesis that the activation of the Ub/proteasome pathway is confined to neoplastic growth.

The expression of multiple ubiquitin-conjugating (E2) enzymes is increased in primary breast cancer, but not in colon cancer. To advance these studies, we also examined the levels of other Ub-conjugating enzymes, UbcH1 and UbcH2, in a subset of breast cancer tissue samples. Consistent with the results described above, all specimens that showed higher levels of UbcH5 also displayed increased levels of UbcH1 and UbcH2 (Fig. 3A). For instance, specimens 1, 3, and 6 showed higher proteasome activity (Fig. 1A), and increased expression of Rpt1 (Fig. 1B), UbcH5 (Fig. 2B), UbcH1, and UbcH2 (Fig. 3A). Specimen 5 showed moderate increase in the levels of E2 enzymes and proteasome subunits. Specimens 2 and 4 did not display increased proteasome activity, or higher levels of proteasome subunits and UbcH1/UbcH2. We also examined the expression of the same set of E2 enzymes in primary colon cancer tissue extracts and did not observe a consistent pattern of expression that would indicate that their levels were coordinately increased (Fig. 3A , bottom).

A well-characterized cell culture model system of breast cancer does not display increased activity of the ubiquitin/proteasome pathway. We anticipated that a cell culture–based breast cancer model would provide an expeditious way to define the underlying role of the Ub/proteasome pathway in cancer. One system that has been recognized as a convincing model for breast cancer involves the MCF10 cell line. Numerous derivatives of this cell line have been generated and reported to display varying degrees of tumorigenicity (23, 24, 27, 28). We characterized a set of cell lines that have been used to model breast cancer (MCF10A, MCF10AT1, MCF10DCIS.COM, and MCF10CA1α). Total protein extracts were prepared, and proteasome activity and abundance were measured (Fig. 3B). We determined that all four cell lines, which represented an untransformed strain, and three distinct lineages with unique tumorigenicity properties, displayed similar proteasome activity (Fig. 3B). Furthermore, immunoblotting analysis showed that the levels of proteasome subunits (Rpn2, Rpt1, and 20S α-subunits) were similar in all the cell lines (Fig. 3C). Immunoblots were also incubated with antibodies against UbcH5 and eEF1A, and as noted above, no difference in expression levels was detected (data not shown). These surprising results underscore the importance of examining primary cancer tissue samples for defining the biochemical changes that accompany breast cancer. Although others have reported that MCF10-derived cells display properties consistent with transformed cells (24), none of the biochemical alterations we observed in primary tissues were detected in this cell culture model. However, conditions of stress such as anoxia, might induce such an effect. A recent study found that the levels of BRCA1 were similar in both growth-arrested and rapidly proliferating MCF10A cells, leading to the conclusion that this tumor suppressor might not control cell cycle progression in breast cancer cells (29). An alternative interpretation, based on the studies described here, suggest that this cell culture–based model system might not reflect all the biochemical changes that underlie breast cancer progression in primary tissues.

High levels of eEF1A are detected in breast cancer. Recent studies showed that 30% to 50% of nascent proteins are degraded before they mature (16–18). Thus, the major fraction of substrates of the proteasome is likely to originate from the ribosome. Cancer cells are metabolically active and are engaged in active protein synthesis. We reported recently that the translational elongation factor eEF1A forms a specific interaction with ubiquitinated nascent proteins and the proteasome in the presence of protein synthesis inhibitors (19). Based on the increased activity of the proteasome, and higher levels of proteasome subunits, Ub-conjugating enzymes and multiubiquitinated proteins in breast cancer tissue samples, we investigated if eEF1A levels were correspondingly increased. Increased expression of eEF1A in breast cancer could suggest the presence of high levels of nascent, damaged proteins. Protein extracts were examined by immunoblotting, and increased levels of eEF1A were detected in most breast cancer specimens (Fig. 4). The increased level of eEF1A in cancers has been described previously (20, 21), and the results shown here suggest that a common cellular response to protein damage might occur during hyperproliferation in breast and other cancers. Although malignant cells are metabolically very active, the dramatically higher levels of eEF1A (30- to 67-fold increased levels in certain specimens) cannot be explained solely by a requirement for increased protein synthesis in metabolically active, neoplastic cells. We propose that the increased abundance of eEF1A reflects its role in the turnover of high levels of nascent damaged proteins. As noted earlier, no change in eEF1A levels was detected in benign solid tumors, confirming our hypothesis that the expression and activity of the Ub/proteasome pathway and its ancillary factors is elevated in primary breast cancer cells.

We used multiple methods to show that the Ub/proteasome proteolytic system is activated in primary human breast cancer tissues. Higher proteasome activity was associated with increased levels of proteasome subunits, Ub-conjugating enzymes (UbcH1, UbcH2, and UbcH5), multiubiquitinated proteins, and the translation elongation factor eEF1A. eEF1A performs a central role in the degradation of nascent, misfolded polypeptide chains, and its elevated expression is consistent with the presence of high levels of damaged proteins in cancer cells. Collectively, these findings indicate that the primary pathway for intracellular protein degradation is activated in breast cancer.

Multiubiquitinated proteins were detected at high levels in cancer cells, although proteasome activity was increased. Metabolically active cancer cells may generate high levels of damaged proteins that are rapidly conjugated to multi-Ub chains. Presumably, this could lead to a compensatory and constitutive increase in basal proteasome activity. In the studies described here, we specifically monitored the activity of the 20S particle. However, a deficiency in 19S function could cause a failure to degrade substrates (and lead to high levels of multi-Ub proteins), even in the presence of functional 20S particles. Additionally, high levels of damaged proteins might exceed the proteolytic capacity of the Ub/proteasome system, leading to both increased proteasome activity and abundance of multi-Ub substrates.

Numerous substrate-targeting pathways channel proteins to the proteasome for degradation. Consequently, increased proteasome activity in cancer cells could reflect a general response to hyperproliferation and stress. Distinct E2 and E3 factors mediate the cellular response to different environmental stresses and growth conditions by ubiquitinating particular substrates. In contrast to breast cancer, we did not observe any increase in the levels of UbcH1, UbcH2, and UbcH5 in colon cancer, suggesting that this set of E2 enzymes might play a key role in the degradation of substrates in breast cancer. We propose, therefore, that the identification of particular E2 and E3 enzymes that channel substrates into the Ub/proteasome pathway can provide important information regarding the cellular response in different cancers.

Our findings reveal a striking consistency; tissue extracts displaying high proteasome activity typically also expressed higher levels of proteasome subunits, Ub-conjugating enzymes, eEF1A, and multiubiquitinated proteins (Fig. 5). Taken together, 90% to 95% of the breast cancer samples were successfully identified by using multiple parameters. Conversely, specimens that did not exhibit a significant change in proteasome activity showed no change in the levels of these targeting factors (sample 4 and two benign solid tumors samples). The ∼3-fold range in proteasome activity in control tissues shows considerable variability in individual basal activity. However, proteasome activity varied by >12-fold in breast cancer specimens, and these values were on average 5-fold higher than the control tissues. The broad range in proteasome activity in the breast cancer specimens could reflect differences in the cellular origin of the malignancy or the state of advancement of the cancer. By standardizing the data derived from the individual samples, to the values detected in their respective patient-matched control tissues (normal adjacent tissue), we readily detected changes in the activity of this proteolytic system.

There is evidence for a direct role for the Ub/proteasome pathway in breast cancer, as the BRCA1 and BRCA2 susceptibility factors have both been linked to this proteolytic system. BRCA1 contains a RING domain that binds Ub-conjugating enzymes and provides Ub-(E3) ligase activity (30). The RING domain in BRCA1 can also bind a deubiquitinating enzyme, BAP1, and both proteins are coexpressed and colocalized (31). Mutations in the RING domain of BRCA1 abolished its ability to function as an E3 ligase, and also blocked its interaction with BAP1, resulting in the loss of tumor suppressing properties of BRCA1. The proteasome inhibitor, Velcade, has demonstrable efficacy in breast cancer (32, 33), consistent with the activation of the Ub/proteasome pathway described here. However, proteasome inhibitors block the terminal step of a proteolytic cascade that receives substrates from numerous physiologic pathways. In contrast, the therapeutic targeting of upstream components, such as E2 and E3 enzymes, might serve to significantly improve the specificity of therapeutic intervention.

In summary, we report that activation of the proteasome, and high-level expression of certain Ub-conjugating enzymes in breast cancer, is consistent with a requirement for proteolysis in hyperproliferating cells. The elevated level of the translation elongation factor eEF1A indicates a failure in the fidelity of protein synthesis, or a defect in the surveillance mechanism that regulates accurate folding of nascent proteins. eEF1A levels increased >50-fold in certain breast cancer specimens, suggesting that a major source of damaged proteins in neoplastic cells may be due to a defect in protein synthesis quality control.

Grant support: NIH grant CA83875 (K. Madura) and Cancer Institute of New Jersey.

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.

We thank the members of the laboratory for comments on the manuscript and Dr. N. Suh (Rutgers University, Piscataway, NJ) for providing the MCF10A set of cell lines.