The perinucleolar compartment (PNC) is a multicomponent nuclear structure enriched with RNAs transcribed by RNA pol III and RNA binding proteins. Studies in cultured cells showed an association between PNC and transformed phenotype. To evaluate the relationship between structure and malignancy in vivo, we examined PNC prevalence (the percentage of cells containing at least one PNC) in normal and cancerous paraffin-embedded breast tissues using immunohistochemistry against a PNC-associated protein. Five hundred nuclei in the most active area of each sample were scored for PNC prevalence. The results show that PNC prevalence significantly correlates with the progression of breast cancer (by the criteria of staging). PNC prevalence in primary tumors, lymph nodes, and distant metastases shows a stepwise increase from a median of 23% in primary tumors to ∼100% in distant metastases. In addition, univariate and multivariate (controlling for tumor size and grade) analyses show that early-stage patients with invasive ductal carcinomas containing a higher PNC prevalence have a significantly poorer prognosis. These findings link PNC prevalence with the progression of breast cancer in vivo and suggest that PNC-containing cells have metastatic advantages. These findings also show the potential of PNC prevalence as a prognostic marker for breast cancer.

Breast cancer is a widespread disease, affecting ∼1 in 10 women during their lifetime (1, 2), making the understanding and treatment of breast cancer an important public health issue. Although early detection and improved treatment have significantly enhanced the survival rate of breast cancer patients, current prognostic and predictive markers are often inaccurate in forecasting the outcome of individual patients (3, 4). The most reliable prognostic markers continue to be the involvement of lymph nodes, the size, and the architectural features of cancer tissues and histologic grading (5). For node-negative patients, there are few prognostic markers that are clinically useful; consequently, these patients routinely undergo adjuvant therapy despite estimates that only 25% of patients benefit from the therapy (4). Many current tumor markers are of limited prognostic value, most likely due to the heterogeneity of oncogenesis. The development of malignancy is a complex process involving extensive molecular and structural alterations, with the pathway of development varying from one tumor to another. The identification of markers that can comprehensively assess malignancies arising from a variety of oncogenic pathways is therefore actively being pursued to improve the accuracy of forecasting individual patient outcomes.

The nucleus, the cellular compartment where genetic information is stored and duplicated and gene expression is regulated at transcriptional and post-transcriptional levels, often undergoes significant structural alterations during malignant transformation (6–8). Histologic grading is to a large extent based on the changes in overall nuclear morphometry, including size, shape, texture, and nucleolar morphology in addition to mitotic index (6, 9). However, most of these changes are difficult to quantitate due largely to interobserver variability. Here, we report a distinct multicomponent nuclear structure, the perinucleolar compartment (PNC), the formation of which is linked to transformation and has the potential to be a less ambiguous cancer biomarker.

The PNC is an irregularly shaped, electron-dense structure that is preferentially associated with the nucleolus (10). It is predominantly present in transformed cells and is rarely detected in primary normal cells in culture. The components of the PNC described thus far include several small RNAs transcribed by RNA pol III (RNase P H1 RNA, RNase MRP RNA, hY RNAs, SRP RNA, and Alu RNA; refs. 11–13) and several RNA binding proteins, including CUG-BP/hNab50, KRSP, Raver1, Rod 1 (14–16), and PTB (polypyrimidine tract binding protein), a protein involved in pre-mRNA splicing, polyadenylation, and translation (17). Transfection studies using wild-type and mutant PTB showed that the RNA binding ability of PTB is required for its PNC localization and that PTB rapidly shuttles in and out of the PNC (14). Pulse labeling using bromouracil for 5 minutes showed that the PNC is enriched with newly synthesized RNA (14). Further studies showed that the structural integrity of the PNC is dependent on the continued production of RNA pol III transcripts and the expression of PTB (13). Together, these observations suggest that the PNC is a dynamic nuclear structure and is involved in RNA metabolism, particularly in transformed cells (18).

Its association with the transformed phenotype in cultured cells and involvement in RNA metabolism suggest that the PNC may result from malignant transformation and may also play a role in the maintenance and/or progression of cancer. To evaluate the association of PNCs with cancer in vivo, we examined PNC prevalence (the percentage of cells that contains at least one PNC) in over 300 normal and cancerous breast tissue samples and statistically analyzed its association with the progression of the disease. Herein we report that the PNC, a multicomponent nuclear structure, is associated with the metastatic potential of breast cancer cells. Its prevalence provides prognostic information additional to that provided by current markers, particularly for node-negative patients.

Tissue Samples. Paraffin-embedded breast tissue samples (n = 301) were obtained from four sources: Dr. Ann D. Thor (formerly of Evanston Northwestern Healthcare), Dr. Elizabeth L. Wiley (Northwestern University Medical School), Comparative Human Tissue Network Midwestern Division (Ohio State University), and Dr. Paula Kovarik (John H. Stroger, Jr., Hospital of Cook County, Chicago, IL) as approved by our institutional and related review boards. Cancerous samples were grouped using the American Joint Committee on Cancer tumor-node-metastasis staging scheme. Patient age ranged from 26 to 88 years with a mean age of 57 years (median, 55.5). Some of the patients with primary infiltrating ductal carcinoma (n = 154) had corresponding clinical and follow-up data including recurrence and distant metastases. Tumor size ranged from 0.5 to 16.0 cm with a mean tumor size of 2.9 cm (median, 2.4). Lymph node status was determined by nodal dissection and H&E-stained histologic sections. Follow-up was calculated from the date of surgery to the last recorded follow-up date (mean follow-up of 7 years with a range of 1-14 years; median, 7.7 years). Local recurrence and distant metastasis were calculated from the date of surgery to the first documented failure date. Tumors were graded using the Nottingham combined histologic grading scheme (9).

Immunohistochemistry. PNCs were detected using a monoclonal antibody (SH54) that specifically recognizes PTB (10), a RNA binding protein that is highly concentrated in the PNC and that serves as a marker for the PNC. Immunohistochemical techniques with antigen retrieval were used on paraffin-embedded tissue sections (4-5 μm) for optimal antibody-antigen interactions (19). The protocol involved deparaffinization in xylene and graded alcohols followed by microwave retrieval for 2 minutes in 10mmol/L citric buffer (pH 6.0) before incubation with the primary and avidin-conjugated secondary antibody. Signals were detected using the biotin-conjugated enzyme horseradish peroxidase (Vector Laboratories, Inc., Burlingame, CA), which converts the chromogen 3,3′-diaminobenzidine (Sigma, St. Louis, MO) into a dark precipitate in the presence of its substrate hydrogen peroxide (Sigma). Tissues were observed on Nikon Eclipse E800 microscope and monochromatic images acquired with a SenSys cooled CCD camera (Photometrics, Tucson, AZ) using MetaView version 4.5 software (Universal Imaging Corp., West Chester, PA). At least 500 nucleiper epithelial cell subtype (e.g., benign epithelium or invasive ductal carcinoma) at most active areas were analyzed for the presence or absence of PNCs. PTB is normally diffusely distributed in the cell nucleus. Any aggregated PTB labeling that is at least 2-fold stronger than the diffused nuclear labeling in the same cell was scored PNC positive. Histopathologic assessment of most active areas for each specimen was independently provided by pathologist collaborators before PNC prevalence scoring.

Statistical Analysis. Pearson product moment correlations, expressed as continuous variables (0-100%), were used to describe the associations between PNC prevalence and clinical and pathologic variables. Unpaired ttests were also expressed as continuous variables (0-100%) to describe the association between PNC prevalence and lymph node status. For univariate survival analyses, factors were separated by dividing data into two or three groups, and odds ratios and confidence levels were calculated. The log-rank statistic was used to determine significance. Survival analyses compared the prognostic value of PNC prevalence. The outcomes selected for the study included disease-free survival (defined as time to local recurrence or distant metastasis) and overall survival (defined as time to death from any cause). The Cox proportional hazards model was used to calculate both univariate and multivariate survival (20).

For multivariate survival, we first constructed models that included all factors univariately associated with survival (excluding PNC prevalence). We then removed the variables one at a time based on the Wald statistic until only statistically significant variables remained. The PNC prevalence variable was added to the models as either a continuous or a dichotomized variable using a cut point of 0 versus ≥1% immunostaining.

The number of positive lymph nodes was logarithmically transformed (after adding the constant 1 to avoid taking log of 0 for node-negative cases), because a plot of Martingale residuals against the number of positive nodes suggested that this transformation gave a better linear relationship between risk and nodes. Tumor size and patient age were entered as continuous variables, and biochemical estrogen receptor levels were dichotomized with a cutoff point of 3 fmol/mg protein or ≥20% of tumor cells positive by immunohistochemistry (institutional cut point).

PNC Prevalence Increases in Parallel with the Clinical Progression of Breast Cancer. Our previous studies with cultured cells derived from a variety of tissues have shown that the PNC is predominantly present in transformed cells (10). We were interested in determining whether the heterogeneity in PNC prevalence between different transformed cell lines reflects differences in aggressiveness of the cell lines. Therefore, we examined a series of breast cancer cell lines with varying degrees of malignancy, including MDA-MB-157, MDA-MB-468, MDA-134-VI, MDA-MB-453, and Hs578T. The results showed that there is no direct correlation between PNC prevalence and the number of chromosomes or the number of passages in culture. However, a higher PNC prevalence was associated with cell lines that induce tumors in nude mice compared with those that do not induce tumors (see Supplemental Data). These findings led to the hypothesis that PNC prevalence may reflect the degree of cellular transformation.

To test this hypothesis in vivo, we examined normal and cancerous breast tissues for PNC prevalence and analyzed PNC prevalence against various clinicopathologic data, including staging and grading. Samples were divided into seven distinct categories. These included breast tissue without pathologic change (henceforth called “normal”), benign hyperplasia, carcinoma in situ (all ductal carcinoma in situ), and invasive carcinomas (Fig. 1A). The invasive carcinomas were further divided into three groups based on the state of lymph node invasion: no lymph node involvement, 1 to 3 lymph nodes involved, and ≥4 lymph nodes involved (Fig. 1A). The seventh category was composed of samples from distant metastases that developed subsequently to the primary diagnosis and adjuvant treatment (Fig. 1A).

Figure 1.

Lower left, immunohistochemical staining of normal and cancerous breast tissue samples at different stages of tumor development. PNCs were detect ed using SH54 monoclonal antibody in paraffin-embedded tissue samples. Arrows, cells containing PNCs. A, low magnification of the samples, demonstrating tissue morphology and architecture. Bar, 20 μm. B, higher magnification zooming of samples in A with better visualization of the PNC-containing cells (arrows). Bar, 5 μm. C, comparisons of PNC prevalence for various diagnostic subgroups.

Figure 1.

Lower left, immunohistochemical staining of normal and cancerous breast tissue samples at different stages of tumor development. PNCs were detect ed using SH54 monoclonal antibody in paraffin-embedded tissue samples. Arrows, cells containing PNCs. A, low magnification of the samples, demonstrating tissue morphology and architecture. Bar, 20 μm. B, higher magnification zooming of samples in A with better visualization of the PNC-containing cells (arrows). Bar, 5 μm. C, comparisons of PNC prevalence for various diagnostic subgroups.

Close modal

PTB, a heterogeneous nuclear ribonucleoprotein that is highly concentrated in the PNC, was used as a marker for PNC detection. An anti-PTB monoclonal antibody (SH54) that we developed (10) was used to immunohistochemically label PNCs in paraffin-embedded breast tissues (n = 301) using an antigen retrieval protocol (see Materials and Methods). Aminimum of 500 epithelial cell nuclei in the most active area (areas with cells of highest grade) for each specimen were analyzed for the presence or absence of PNCs. The slides were not counterstained. Any nucleus containing aggregates with PTB labeling intensity 2-fold or stronger than the diffuse nucleoplasmic staining was scored PNC positive. The measurement of fold differences in labeling intensity within the same nucleus provides a self-controlled condition. In addition, positive and negative controls were simultaneously examined with each batch of labeling. The scoring process was done in a blinded fashion such that the examiner did not have prior knowledge of the patient information, including tumor-node-metastasis staging, grading, or patient outcome.

Observations of the tissues using lower magnification allowed the visualization of general tissue morphology and architecture (Fig. 1A), whereas higher magnification was used to more clearly discern the PNC-containing cells (Fig. 1B). The immunostaining patterns of PTB in tissues derived from different stages of neoplasia ranging from benign growths to distant metastases exhibited significant morphologic differences (Fig. 1A and B). Normal breast epithelia showed primarily diffused labeling of PTB in the nucleus and did not show detectable PNCs (Fig. 1A, and B, top left). Apocrine metaplasias showed a mean PNC prevalence of 5.4%, consistent with its more benign nature (Fig. 1A, and B, middle top, and C). Ductal carcinomas in situ and invasive ductal carcinomas without lymph node involvement had a mean PNC prevalence of 13.6% and 18.5%, respectively (Fig. 1A, and B, top right and bottom left, and C). Invasive ductal carcinomas with either 1 to 3 or ≥4 lymph nodes involved showed a mean PNC prevalence of 21% and 32%, respectively (Fig. 1A, and B, middle bottom, and C). In distant metastases, the majority of cancer cells (95%) exhibited a punctuate PTB staining pattern in the nucleus, which we interpret as the presence of multiple PNCs (Fig. 1A, and B, lower right, and C). The differences in the mean PNC prevalence between any two categories were significant (P < 0 0.001; except between ductal carcinoma in situ and invasive carcinoma without lymph node involvement, P < 0.05) as determined by Student's unpaired t tests (Table 1A). These results show that PNC prevalence increases in parallel with the malignant progression of breast cancer in vivo, consistent with the in vitro findings in breast cancer derived cell lines.

Table 1.

Statistical analysis of data

A. Unpaired Student's t test correlation between different stages of breast cancer
NormalBenignIn situInvasive node negativeInvasive node positiveMetastatic
n 21 31 10 61 93 39 
Benign <0.0001 — <0.0001 <0.0001 <0.0001 <0.0001 
In situ <0.0001 <0.0001 — 0.0276 <0.0001 <0.0001 
Invasive node negative <0.0001 <0.0001 0.0276 — <0.0001 <0.0001 
Invasive lymph node positive <0.0001 <0.0001 <0.0001 <0.0001 — <0.0001 
Metastatic
 
<0.0001
 
<0.0001
 
<0.0001
 
<0.0001
 
<0.0001
 

 
        
B. Pearson product moment correlations between PNC and pathologic variables for invasive carcinoma       
 n R (95% CI) P    
Patient age 138 −0.139 (−0.30 to 0.03) 0.1    
Tumor size T1, T2, T3 149 0.289 (0.13-0.43) 0.0003    
Lymph node 154 0.522 (0.40-0.63) <0.0001    
Grade 135 0.206 (0.04-0.36) 0.00162    
Estrogen receptor 140 0.099 (−0.07 to 0.26) 0.24    
Progesterone receptor 132 −0.081 (−0.25 to 0.09) 0.36    
A. Unpaired Student's t test correlation between different stages of breast cancer
NormalBenignIn situInvasive node negativeInvasive node positiveMetastatic
n 21 31 10 61 93 39 
Benign <0.0001 — <0.0001 <0.0001 <0.0001 <0.0001 
In situ <0.0001 <0.0001 — 0.0276 <0.0001 <0.0001 
Invasive node negative <0.0001 <0.0001 0.0276 — <0.0001 <0.0001 
Invasive lymph node positive <0.0001 <0.0001 <0.0001 <0.0001 — <0.0001 
Metastatic
 
<0.0001
 
<0.0001
 
<0.0001
 
<0.0001
 
<0.0001
 

 
        
B. Pearson product moment correlations between PNC and pathologic variables for invasive carcinoma       
 n R (95% CI) P    
Patient age 138 −0.139 (−0.30 to 0.03) 0.1    
Tumor size T1, T2, T3 149 0.289 (0.13-0.43) 0.0003    
Lymph node 154 0.522 (0.40-0.63) <0.0001    
Grade 135 0.206 (0.04-0.36) 0.00162    
Estrogen receptor 140 0.099 (−0.07 to 0.26) 0.24    
Progesterone receptor 132 −0.081 (−0.25 to 0.09) 0.36    

PNC Prevalence Positively Correlates with the Clinicohistologic Classifications of Breast Cancer and Inversely Correlates with Patient Outcome. Pearson moment correlation analyses were done to explore correlations between PNC prevalence and currently used clinical markers. PNC prevalence correlates with tumor size (P = 0.0003), node involvement (P < 0.0001), and histologic tumor grade (P = 0.00162; Table 1B). Although no significant association was observed between PNC prevalence and patient's age, estrogen receptor, or progesterone receptor status (Table 1B), its correlations with several clinically important markers further show that PNC prevalence is indicative of the trends of breast cancer progression.

To evaluate the relationship between PNC prevalence and patient outcome, univariate survival analyses were done. PNC prevalence, patient age, tumor grade, and tumor size were each significantly associated with disease-free survival (Table 2). PNC prevalence, tumor grade, and tumor size were significantly associated with overall survival (Table 2). Using either the median of all carcinomas (18%) or that of invasive carcinomas (23%), PNC prevalence significantly correlates with poor patient outcome (Table 2). To further determine whether PNC prevalence contains prognostic information under well-controlled conditions, we compared PNC prevalence between samples from matched patients with similarly diagnosed invasive breast carcinoma, the disease of which either did or did not progress. Samples (n = 129) were divided into three categories based on nodal involvement and follow-up information. Univariate analyses showed that tumors from patients who developed recurrence or metastases showed significantly greater PNC prevalence compared with those from patients, the cancer of which did not progress (Fig. 2A). These differences can be observed in both node-negative (P < 0.0001) and 1 to 3 node-positive (P < 0.0128; Fig. 2A) subgroups. However, PNC prevalence did not significantly correlate with the outcome of patients in the group with ≥4 nodes involved (P < 0.0304; Fig. 2A). The lack of a significant correlation could be due to the enrichment of cancer cells of advanced transformation states in these late-stage tumors. Kaplan-Meier analyses, using the median PNC prevalence (23%) for invasive carcinomas as a cutoff value, graphically showed these survival associations over median follow-up of 7.7 years. Patients with node-negative disease, but high PNC prevalence (≥23%) had a significantly poorer disease-free survival than did similar patients with low PNC prevalence (<23%; P = 0.0013; Fig. 2B). These findings indicate that PNC prevalence positively correlates with the relapse of breast cancer, particularly those of patients diagnosed as early-stage invasive breast carcinomas. These results further support that the formation of the PNC associates with the aggressive behavior of tumor cells and show that PNC prevalence contains prognostic value for patients with invasive carcinomas, particularly for those with negative nodes.

Table 2.

Univariate analysis for factors associated with survival

FactorsnDisease-free survival
Overall survival
Odd ratios (95% CI)POdd ratios (95% CI)P
Age (y)      
<50 52 0.0009 0.98 
≥50 77 0.378 (0.21-0.67)  1.006 (0.61-1.65)  
Size (cm)      
≥2 61 0.0105 0.0092 
68 2.336 (1.22-4.48)  2.038 (1.19-3.48)  
Node negative 59 0.31 0.39 
Node positive 70 1.34 (0.76-2.36)  1.237 (0.76-2.01)  
PNC      
<18.6 46 0.0007 0.013 
≥18.6 83 4.018 (1.80-8.97)  2.048 (1.16-3.61)  
<23.6 64 0.0113 0.0113 
≥23.6 65 1.907 (1.16-3.14)  1.907 (1.16-3.14)  
Elston grade      
46 0.0011 0.0054 
38 2.568 (1.07-6.15)  2.273 (1.14-4.53)  
45 3.944 (1.78-8.77)  2.855 (1.50-5.42)  
FactorsnDisease-free survival
Overall survival
Odd ratios (95% CI)POdd ratios (95% CI)P
Age (y)      
<50 52 0.0009 0.98 
≥50 77 0.378 (0.21-0.67)  1.006 (0.61-1.65)  
Size (cm)      
≥2 61 0.0105 0.0092 
68 2.336 (1.22-4.48)  2.038 (1.19-3.48)  
Node negative 59 0.31 0.39 
Node positive 70 1.34 (0.76-2.36)  1.237 (0.76-2.01)  
PNC      
<18.6 46 0.0007 0.013 
≥18.6 83 4.018 (1.80-8.97)  2.048 (1.16-3.61)  
<23.6 64 0.0113 0.0113 
≥23.6 65 1.907 (1.16-3.14)  1.907 (1.16-3.14)  
Elston grade      
46 0.0011 0.0054 
38 2.568 (1.07-6.15)  2.273 (1.14-4.53)  
45 3.944 (1.78-8.77)  2.855 (1.50-5.42)  

NOTE: All values were calculated using continuous variable. The median PNC prevalence was 18.6 for all cases tested and 23.6 for primary invasive cases with follow-up. When node status was used as a continuous variable, the χ2 method was used to determine statistical significance.

Figure 2.

A, comparisons of PNC prevalence of samples at primary diagnosis from patients with invasive breast carcinoma who did or did not develop relapse in a form of distant metastases. Samples were subgrouped based on nodal involvement. B, Kaplan-Meier survival curves for node-negative patients using PNC prevalence data. Median patient follow-up time was 7.7 years.

Figure 2.

A, comparisons of PNC prevalence of samples at primary diagnosis from patients with invasive breast carcinoma who did or did not develop relapse in a form of distant metastases. Samples were subgrouped based on nodal involvement. B, Kaplan-Meier survival curves for node-negative patients using PNC prevalence data. Median patient follow-up time was 7.7 years.

Close modal

PNC Prevalence Provides Independent Prognostic Information for Early-Stage Breast Cancer Patients. To determine the contribution of PNC prevalence to prognosis in the context of other prognostic markers, we did multivariate analyses. The addition of the PNC prevalence of primary tumors to other markers for all patients, including those having node-negative and node-positive invasive carcinomas, improved the accuracy of disease-free survival prediction (Table 3A). However, the addition of PNC prevalence in these analyses did not significantly improve overall survival prediction (Table 3A).

Table 3.

Cox proportional hazards model of multivariate survival

FactorPatients (n)Events (n)χ2Δχ2P
A. All patients      
    Disease-free survival      
        Age + size 124 45 21.244 6.056 0.0139 
        Age + size + PNC 124 45 27.3   
    Overall survival      
        Size 124 61 15.198 1.382 0.24 
        Size + PNC 124 61 16.58   
B. Lymph node-negative patients      
    Disease-free survival      
        Grade 56 20 3.398 11.742 0.0048 
        Grade + PNC 56 20 15.14   
    Overall survival      
        Size + grade 56 27 8.811 4.803 0.0284 
        Size + grade + PNC 56 27 13.614   
FactorPatients (n)Events (n)χ2Δχ2P
A. All patients      
    Disease-free survival      
        Age + size 124 45 21.244 6.056 0.0139 
        Age + size + PNC 124 45 27.3   
    Overall survival      
        Size 124 61 15.198 1.382 0.24 
        Size + PNC 124 61 16.58   
B. Lymph node-negative patients      
    Disease-free survival      
        Grade 56 20 3.398 11.742 0.0048 
        Grade + PNC 56 20 15.14   
    Overall survival      
        Size + grade 56 27 8.811 4.803 0.0284 
        Size + grade + PNC 56 27 13.614   

NOTE: All values were calculated using continuous variable, except, progesterone receptor, which was calculated as a negative/positive value.

When we analyzed the contribution of PNC prevalence to survival prediction in patients grouped according to their nodal involvement, the most striking contribution of PNC prevalence to survival prediction is observed for stage I node-negative patients. In these patients, tumor grade for disease-free survival and tumor size for overall survival were independently associated with patient outcome. Adding PNC prevalence to tumor grade significantly improved survival prediction for disease-free survival (Table 3B) and adding PNC prevalence to both tumor grade and size significantly improved survival prediction for overall survival (Table 3B). However, addition of PNC prevalence to these markers for node-positive patients did not improve survival prediction (data not shown), consistent with the results of univariate analyses showing that PNC prevalence offers little prognostic information for advanced-stage patients. Together, our observations indicate that PNC prevalence provides additional prognostic information that can improve the accuracy for breast cancer prognosis, particularly for patients with stage I node-negative diseases.

PNC Prevalence Reflects Growth and Metastatic Potential of Tumor Cells. PNC prevalence gradually increases in primary tumors of progressing stages, culminating at ∼95% in the distant metastases (see Introduction). This finding raised the question of whether the presence of the PNC reflects growth and/or metastatic advantages of tumor cells. To address this question, we compared the PNC prevalence in the primary breast tumor with that in lymph node metastases from the same patients. Our rationale was that at least two steps are involved in metastases, entering the circulation and penetration into the tissue and reseeding at distant sites. Both steps likely represent selective processes for cells with metastatic capability. The breast cancer cells in lymph node samples therefore should be those at an intermediate step, en route to distant organ metastases, and thus may exhibit molecular and cellular characteristics that are intermediate between those of original tumors and distant metastases (21). If the PNC is indeed indicative of metastatic behavior, we would expect an enrichment of PNC-containing cells in lymph node invasions (representing cells that passed the first step) and PNCs being even more enriched in distant metastases (representing cells that passed both steps). As we have shown in Fig. 1 that PNC prevalence in distant metastases is ∼100%, we were interested in comparing the PNC prevalence in lymph node invasions with their primary tumors. To control for the interpatient variability, the primary tumor and affected lymph nodes were obtained from the same patients and PNC prevalence was scored from the same assay for each paired sample. We examined a total of 23 pairs of samples that did not overlap with those described and analyzed in this study.

The results showed that there was a statistically significant difference between PNC prevalence in primary breast tumors (47.9%) and that in lymph node metastases (76.3%) and those in distant metastasis (95%; Fig. 3A). The average PNC prevalence is higher than those in Fig. 1C, because in these analyses only grades 2 and 3 tumors were evaluated due to the general lack of node involvement for lower-grade tumors. In addition, patients with grade 3 breast carcinomas, as determined by the Bloom-Richardson grading system (9), showed a higher PNC prevalence in both primary tumors and lymph nodes compared with patients with grade 2 tumors (Fig. 3B). These results show a significant enrichment of PNC-containing cells in lymph node invasions and ∼100% enrichment in distant metastases, demonstrating a successive selection of PNC-containing cells for metastases. These observations further support the idea that PNC-containing cells may have a growth and metastatic advantages over those without detectable PNCs.

Figure 3.

A, comparisons of PNC prevalence among primary tumors, lymph node invasions, and distant metastases. Samples for primary tumor and lymph node are paired and derived from the same patients. B, comparisons of PNC prevalence between primary tumors and paired lymph node invasions in subgroups based on their tumor grades.

Figure 3.

A, comparisons of PNC prevalence among primary tumors, lymph node invasions, and distant metastases. Samples for primary tumor and lymph node are paired and derived from the same patients. B, comparisons of PNC prevalence between primary tumors and paired lymph node invasions in subgroups based on their tumor grades.

Close modal

Our studies show that PNC prevalence correlates with breast epithelial transformation both in vitro and in vivo. PNC prevalence in primary tumors increases in parallel with disease progression as defined by clinical staging. There also is a successive enrichment of PNC-containing cells from primary tumors to involved nodes to distant metastases. In addition, patients with tumors of lower PNC prevalence have significantly better long-term survival than those with tumors of high PNC prevalence. These findings together suggest that the presence of the PNC, a multicomponent cellular structure, reflects the advanced stages of malignant transformation and that cells containing PNCs have growth and metastatic advantages.

Although PNC prevalence positively correlates with tumor grade, size, and nodal involvement, it also has independent prognostic value beyond these routinely used clinical factors, indicating its potential to be a useful prognostic marker. In current clinical practice, the size of the tumor, the number of lymph nodes involved, and the metastatic status are the classic factors that are used to stage breast cancers and to guide therapeutic strategies for breast cancer patients. In general, survival rates decrease with more advanced stages. Lymph node status is the most significant predictor for survival, and chances of survival are inversely related to the number of affected lymph nodes at the time of diagnosis (22). For node-negative patients, tumor size and histologic grade (based on the modified Bloom-Richardson system that uses tubule formation, nuclear features, and proliferation rates; ref. 9) are the best currently used predictors of patient outcomes (6). Although these tumor characteristics are useful for both diagnosis and prognosis, they provide limited information for the prediction of the breast cancer progression. For example, ∼25% of node-negative patients develop relapses (4), yet there are no markers that can accurately identify this group of patients.

Further compounding an already complex situation is the fact that breast cancer is a heterogenous grouping of diseases of many subtypes. At the cellular level, a single breast tumor generally contains a heterogenous population of cells. Even morphologically similar breast cancers may have inherent differences in their genetic and molecular composition, leading to differences in patient responses to treatments and disease relapse (3, 23, 24). The growth of a tumor generally allows for selection of the most aggressive cells during the expansion of its population, eventually leading to metastasis and the accompanying poor prognosis. Therefore, an ideal marker for prognosis should be one that is indicative of the number of highly aggressive cancer cells at the original lesion at the time of diagnosis. Our finding that the prevalence of the PNC, a multicomponent structure, positively associates with breast cancer progression and inversely correlates with patient survival indicates a potential for PNC prevalence to be such a marker. This potential is further supported by the finding that PNC prevalence provides prognostic information additional to the standard markers, particularly for stage I breast cancer patients. The representation of PNC prevalence for breast malignancy in part may be due to the nature of the PNC itself. Because the PNC is a multicomponent nuclear structure rather than a single gene or gene product, its formation may represent the interplay of multiple cellular processes and is more likely to reflect the advanced transformation derived from changes of heterogeneous pathways. A larger-scale preclinical investigation will be conducted to confirm the prognostic value of PNC prevalence in breast cancer and establish the basis for future clinical utilization.

A growing number of molecular and immunocytochemical markers are being used or are in trial to provide additional information for breast cancer prognosis and prediction (22, 25). The PNC may have distinct advantages over many of these, because the PNC is a structurally distinct multicomponent nuclear domain that reflects an overall aggressiveness of cancer cellular behavior. This is in contrast to any single gene or gene products that may only be altered in a subset of breast malignancies and may provide limited diagnostic and prognostic information. For instance, estrogen receptor activity in breast carcinoma predicts tumor response to estrogen ablation therapy and progesterone receptor activity forecasts better survival (26). Another molecular marker, HER-2, a growth hormone receptor, is associated with the aggressive behavior of breast cancer (27–30). Although these markers are useful for predicting response to specific therapy, their overall prognostic accuracy is limited. For example, a substantial number of patients with tumors that are estrogen receptor positive will develop relapse (31).

More recently, gene expression profiling using microarray techniques shows promise as a test to provide a more comprehensive representation of the metastatic behavior of breast cancer cells. Abnormal expression of 78 to 320 genes is associated with metastases after primary diagnosis (32, 33). Although expression profiling is encouraging, its clinical utility is yet to be defined and clarified in large-scale studies. In addition, expression profiling uses samples that can contain cell types other than cancer cells, whereas PNC prevalence scoring offers resolution at the single cell level that could yield less ambiguous results.

Although the formation of the PNC correlates with the transformed phenotype, the function of the structure and its exact relationship to malignancy remains to be understood. Recent studies from our laboratory showed that the integrity of PNCs is dependent on RNA pol III transcription, suggesting that the PNC may be functionally related to the regulation of RNA pol III gene expression (13). Some of the small RNAs transcribed by pol III also enriched in the PNC are involved in ribosome synthesis and in translation, both of which are significantly altered during transformation (34). As a structure associated with transformed cells and with prevalence that increases with breast cancer progression, the PNC is likely to either be a product of and/or contribute to transformation. Studies are currently under way to investigate the link between PNCs and nucleolar functions such as ribosome synthesis. These investigations will help clarify the roles of PNCs in malignant cells and could present new therapeutic targets for the treatment of breast cancer.

In summary, we have shown that PNC prevalence positively correlates with the disease progression of breast cancer in vivo. PNC prevalence is an independent prognostic variable by multivariate analyses, particularly for early-stage invasive breast carcinomas. These findings suggest that the formation of the PNC is indicative of the aggressive behavior of breast cancer cells and show the clinical relevance of this nuclear structure as an independent prognostic marker for the management of breast cancer patients. Future investigations will seek to verify these results on a larger scale and clarify the specific role of PNCs in transformed cells in vivo.

Grant support: National Cancer Institute grants 5R21 CA84369 and R33 CA097761-02 and Specialized Program of Research Excellence in Breast Cancer for the Robert H. Lurie Comprehensive Cancer Center grant 1 P50 CA89018-01 (S. Huang) and National Cancer Institute grant CA89018, ENH Auxiliary Support of Breast Cancer Research (derived from the Annual American Crafts Exposition), and Departments of Surgery and Pathology, Evanston Northwestern Healthcare Evanston Hospital (A.D. Thor).

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.

1
Eifel P, Axelson JA, Costa J, et al. National Institutes of Health Consensus Development Conference Statement: adjuvant therapy for breast cancer, November 1-3, 2000.
J Natl Cancer Inst
2001
;
93
:
979
–89.
2
Howe HL, Wingo PA, Thun MJ, et al. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends.
J Natl Cancer Inst
2001
;
93
:
824
–42.
3
Chung MA, Wazer D, Cady B. Contemporary management of breast cancer.
Obstet Gynecol Clin North Am
2002
;
29
:
173
–88.
4
Chia SK, Speers CH, Bryce CJ, Hayes MM, Olivotto IA. Ten-year outcomes in a population-based cohort of node-negative, lymphatic, and vascular invasion-negative early breast cancers without adjuvant systemic therapies.
J Clin Oncol
2004
;
22
:
1630
–7.
5
Broders AC Sr. Carcinoma of the breast (including carcinoma in situ) and its grades of malignancy and prognosis.
W V Med J
1953
;
49
:
311
–6.
6
Kamath R, Leary DJ, Huang S. Nuclear components and tumor markers. In: Visions of the nucleus-eukaryotic DNA. Stevenson Ranch (CA): American Scientific Publishers; 2004.
7
Nickerson JA. Nuclear dreams: the malignant alteration of nuclear architecture.
J Cell Biochem
1998
;
70
:
172
–80.
8
Leman ES, Getzenberg RH. Nuclear matrix proteins as biomarkers in prostate cancer.
J Cell Biochem
2002
;
86
:
213
–23.
9
Elston CW, Ellis IO. Pathological prognostic factors in breast cancer. I. The value of histological grade in breast cancer: experience from a large study with long-term follow-up.
Histopathology
1991
;
19
:
403
–10.
10
Huang S, Deerinck TJ, Ellisman MH, Spector DL. The dynamic organization of the perinucleolar compartment in the cell nucleus.
J Cell Biol
1997
;
137
:
965
–74.
11
Matera AG, Frey MR, Margelot K, Wolin SL. A perinucleolar compartment contains several RNA polymerase III transcripts as well as the polypyrimidine tract-binding protein, hnRNP I.
J Cell Biol
1995
;
129
:
1181
–93.
12
Lee B, Matera AG, Ward DC, Craft J. Association of RNase mitochondrial RNA processing enzyme with ribonuclease P in higher ordered structures in the nucleolus: a possible coordinate role in ribosome biogenesis.
Proc Natl Acad Sci U S A
1996
;
93
:
11471
–6.
13
Wang C, Politz JC, Pederson T, Huang S. RNA polymerase III transcripts and the PTB protein are essential for the integrity of the perinucleolar compartment.
Mol Biol Cell
2003
;
14
:
2425
–35.
14
Huang S, Deerinck TJ, Ellisman MH, Spector DL. The perinucleolar compartment and transcription.
J Cell Biol
1998
;
143
:
35
–47.
15
Timchenko LT, Miller JW, Timchenko NA, et al. Identification of a (CUG)n triplet repeat RNA-binding protein and its expression in myotonic dystrophy.
Nucleic Acids Res
1996
;
24
:
4407
–14.
16
Huttelmaier SIS, Grosheva I, Rudiger M, Singer RH, Jockusch BM. Raver1, a dual compartment protein, is a ligand for PTB/hnRNPI and microfilament attachment proteins.
J Cell Biol
2001
;
155
:
775
–85.
17
Wagner EJ, Garcia-Blanco MA. Polypyrimidine tract binding protein antagonizes exon definition.
Mol Cell Biol
2001
;
21
:
3281
–8.
18
Huang S. Review: perinucleolar structures.
J Struct Biol
2000
;
129
:
233
–40.
19
Zukerberg LR, Yang WI, Gadd M, et al. Cyclin D1 (PRAD1) protein expression in breast cancer: approximately one-third of infiltrating mammary carcinomas show overexpression of the cyclin D1 oncogene.
Mod Pathol
1995
;
8
:
560
–7.
20
Cox DR. Regression models and life-tables.
J R Stat Soc Ser B
1972
;
34
:
187
–202.
21
Nathanson SD. Insights into the mechanisms of lymph node metastasis.
Cancer
2003
;
98
:
413
–23.
22
Ross JS, Linette GP, Stec J, et al. Breast cancer biomarkers and molecular medicine.
Expert Rev Mol Diagn
2003
;
3
:
573
–85.
23
Mokbel K. Towards optimal management of ductal carcinoma in situ of the breast.
Eur J Surg Oncol
2003
;
29
:
191
–7.
24
Ottesen GL. Carcinoma in situ of the female breast. A clinico-pathological, immunohistological, and DNA ploidy study.
APMIS Suppl
2003
;
:
1
–67.
25
Fitzgibbons PL, Page DL, Weaver D, et al. Prognostic factors in breast cancer. College of American Pathologists Consensus Statement 1999.
Arch Pathol Lab Med
2000
;
124
:
966
–78.
26
Buzdar AU. Endocrine therapy in the treatment of metastatic breast cancer.
Semin Oncol
2001
;
28
:
291
–304.
27
Mack L, Kerkvliet N, Doig G, O'Malley FP. Relationship of a new histological categorization of ductal carcinoma in situ of the breast with size and the immunohistochemical expression of p53, c-erb B2, bcl-2, and ki-67.
Hum Pathol
1997
;
28
:
974
–9.
28
Moreno A, Lloveras B, Figueras A, et al. Ductal carcinoma in situ of the breast: correlation between histologic classifications and biologic markers.
Mod Pathol
1997
;
10
:
1088
–92.
29
Thor A. HER2—a discussion of testing approaches in the USA.
Ann Oncol
2001
;
12
Suppl 1:
S101
–7.
30
Thor A.
Are patterns of HER-2/neu amplification and expression among primary tumors and regional metastases indicative of those in distant metastases and predictive of Herceptin response? J Natl Cancer Inst
2001
;
93
:
1120
–1.
31
Barker S. Anti-estrogens in the treatment of breast cancer: current status and future directions.
Curr Opin Investig Drugs
2003
;
4
:
652
–7.
32
van't Veer LJ, Dai H, van de Vijver MJ, et al. Gene expression profiling predicts clinical outcome of breast cancer.
Nature
2002
;
415
:
530
–6.
33
Sorlie T, Tibshirani R, Parker J, et al. Repeated observation of breast tumor subtypes in independent gene expression data sets.
Proc Natl Acad Sci U S A
2003
;
100
:
8418
–23.
34
Busch H, Ballal NR, Busch RK, et al. Controls of nucleolar function in cancer cells.
Adv Exp Med Biol
1977
;
92
:
125
–80.