Stathmin expression and its relationship to microtubule-associated protein tau and outcome in breast cancer

Authors


Abstract

BACKGROUND:

Microtubule-associated proteins (MAPs) endogenously regulate microtubule stability. Here, the prognostic value of stathmin, a destabilizing protein, was assessed in combination with MAP-tau, a stabilizing protein, in order to evaluate microtubule stabilization as a potential biomarker.

METHODS:

Stathmin and MAP-tau expression levels were measured in a breast cancer cohort (n = 651) using the tissue microarray format and quantitative immunofluorescence (AQUA) technology, then correlated with clinical and pathological characteristics and disease-free survival.

RESULTS:

Univariate Cox proportional hazard models indicated that high stathmin expression predicts worse overall survival (hazard ratio [HR] = 1.48; 95% confidence interval [CI] = 1.119-1.966; P = .0061). Survival analysis showed 10-year survival of 53.1% for patients with high stathmin expression versus 67% for low expressers (log-rank, P < .003). Cox multivariate analysis showed high stathmin expression was independent of age, menopausal status, nodal status, nuclear grade, tumor size, and estrogen receptor, progesterone receptor, and human epidermal growth factor receptor 2 expression (HR = 1.19; 95% CI = 1.03-1.37; P = .01). The ratio of MAP-tau to stathmin expression showed a positive correlation to disease-free survival (HR = 0.679; 95% CI = 0.517-0.891; P = .0053) with a 10-year survival of 65.4% for patients who had a high ratio of MAP-tau to stathmin versus 52.5% 10-year survival rate for those with a low ratio (log-rank, P = .0009). Cox multivariate analysis showed the ratio of MAP-tau to stathmin was an independent predictor of overall survival (HR = 0.609; 95% CI = 0.422-0.879; P = .008).

CONCLUSIONS:

Low stathmin and high MAP-tau are associated with increased microtubule stability and better prognosis in breast cancer. Cancer 2012. © 2012 American Cancer Society.

Accumulating evidence indicates that microtubule-associated proteins (MAPs) may be associated with varying therapeutic response rates and resistance to taxanes in patients with breast cancer.1–7 MAP-tau is among the best characterized microtubule stabilizing protein that functions to bundle, space, and assemble microtubules.8–11 In contrast, stathmin (stathmin1, OP18, metablastin, p19, prosolin) is a ubiquitous cytosolic phosphoprotein and key regulator of cell division due to its depolymerization of microtubules in a phosphorylation-dependent manner. Proper assembly of the mitotic spindle requires desphosphorylation and inactivation of stathmin before cell entry into mitosis, followed by stathmin reactivation during cytokinesis.12–14 When activated outside of mitosis (eg, during cytokinesis), stathmin regulates dynamic equilibrium of tubulin polymerization through α-β tubulin dimer sequestration and/or promotion of microtubule plus-end catastrophe15–18 and influences microtubule–actin associations.19 The ability of stathmin to remodel microtubule networks through tubulin polymerization and its microtubule–actin associations indicates a role for stathmin in tumor cell migration and invasion.20, 21

The constitutive activation and overexpression of stathmin expression has been associated with a variety of human malignancies, including breast, lung, gastric, ovarian, cervical, prostate, urothelial, hepatocellular, and colorectal carcinomas.22–28 In breast cancer cell lines, stathmin overexpression has been associated with reduced taxane sensitivity and increased resistance to taxane-based chemotherapy due to delayed entry into mitosis.2, 7, 29, 30 In patients with breast cancer, overexpression of stathmin messenger RNA has been correlated with high mitotic index, loss of estrogen receptor (ER) and progesterone receptor (PR) and poor prognosis.31 Upon examination by immunohistochemistry (IHC), high stathmin was associated with phosphatase and tensin homolog (PTEN)-negative tumors and predicted distant metastasis.32

The goal of this study was to assess the expression of stathmin to determine its prognostic value in a large cohort of primary breast cancer patients. Furthermore, because cellular functions require a critical balance between both microtubule stabilizers such as MAP-tau and destabilizers such as stathmin, we hypothesized that a 2-marker model might provide improved prognostic information for patients with breast cancer.

MATERIALS AND METHODS

Patient and Cohort Characteristics

Formalin-fixed paraffin-embedded primary breast cancer tumors resected from 651 patients at Yale University/New Haven Hospital between 1962 and 1983 were obtained from the archives of the Pathology Department, Yale University (New Haven, Conn) and have been previously described in detail.33 Specimens and associated clinical information were collected under the guidelines and approval of the Yale Human Investigation Committee under protocol #8219 to D.L. Rimm. All tissue used in these studies was collected after patient consent or waiver of patient consent (for example, when patients were deceased) in accordance with protocol #8219.

Tissue Microarrays

Tissue microarrays (TMAs) were constructed as described.33 Index TMAs were made from cell pellets from cell lines BT-20, BT-474, MCF-7, MDA-MB-231, MDA-MB-361, MDA-MB-453, MDA-MB-468, SKBR3, ZR-75-1, and CaCo-2 and were formalin-fixed and paraffin-embedded (detailed protocol is described elsewhere34, 35). Pellets were then cored and added to a panel of 40 breast cancer patient controls on a TMA. Index TMAs served as control arrays during antibody validation and immunofluorescence staining to confirm assay reproducibility and for normalization of scores between slides.

Cell Culture

A panel of breast cancer cell lines and several non–breast cancer controls were purchased from ATCC (Manassas, Va) and cultured in order to measure endogenous levels of stathmin and included BT-20, BT-474, MCF-7, MDA-MB-231, MDA-MB-361, MDA-MB-453, MDA-MB-468, SKBR3, ZR-75-1, UAC812, and CaCo-2 cells. Cells were maintained per ATCC instructions. Because cell lines are used as expression controls, they were not authenticated after purchase.

Western Blotting

Whole-cell lysates were prepared, and western blots were performed using standard methods. Bands were quantified with National Institutes of Health ImageJ software and normalized to β-actin. The area under the curve (as measured by ImageJ) versus AQUA (Automated QUantitative Analysis; HistoRx) score for each cell line was plotted, and the linear regression was determined (Fig. 1).

Figure 1.

Stathmin expression in cell line controls and frequency distribution is shown in normal breast tissue and in the Yale University cohort. (A) Stathmin protein as detected by western blot is shown in 11 cell line controls (top panel); CaCo-2 cells were inserted as a positive control, and β-actin served as a loading control. (Bottom panel) The distribution of stathmin AQUA scores are shown from the same cell lines embedded in the breast cancer index array. Note that we were unable to find a cell line that did not express stathmin. (B) Stathmin siRNA knockdown with scrambled control after 24 hours is shown by immunofluorescence of BT-20 cells (upper panel) and western blot from lysates from the same cells. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a loading control. (C) Distribution of average stathmin AQUA scores in the Yale University cohort shows stratification of the cohort using the cutpoint of 25 (dotted arrow) derived from the mean level of expression in normal breast tissue. Inset (light blue bars) shows the distribution of average stathmin AQUA scores in normal breast tissue.

Small Interfering RNA Knockdown

For small interfering RNA (siRNA) transfection, 2 × 105 BT-20 cells were seeded onto 6-well plates in triplicate. The next day, cells were transfected with 100 pmol siRNA duplexes targeting scrambled or stathmin, using Lipofectamine RNAiMax (Invitrogen), and were incubated for 24, 48, and 72 hours. Cells were collected and lysed in sodium dodecyl sulfate (SDS) sample buffer for protein extraction and SDS polyacrylamide gel electrophoresis, and immunoblotting was used to evaluate inhibitory effects. Stealth RNAi siRNA for stathmin (HSS142799) and Stealth RNAi siRNA Negative Control Lo GC (12935-200) were purchased from Invitrogen. In addition, cells were seeded onto coverslips and treated as above for immunofluorescence analysis of stathmin siRNA. Again, cells were incubated for 24, 48, and 72 hours after transfection, then fixed with 4% paraformaldehyde, permeabilized with 0.2% Triton X-100 in phosphate-buffered saline (American Bioanalytical), blocked with 2% bovine serum albumin in phosphate-buffered saline, incubated with stathmin rabbit monoclonal antibody for 1 hour, incubated with Alexa-546 conjugated goat anti-mouse secondary antibody (Invitrogen) for 1 hour, and finally coverslipped with ProLong mounting medium (ProLong Gold, P36931; Molecular Probes) containing 4′,6-diamidino-2-phenylindole.

Antibodies and Immunofluorescence

Yale University cohort TMAs were immunostained using MAP-tau monoclonal antibody, which recognizes all human MAP-tau isoforms (1:750; mouse monoclonal, clone 2B2.100/T1029; US Biological, Swampscott, Mass), and stathmin/op18 monoclonal antibody (1:200,000; rabbit monoclonal, clone EP1573Y; Epitomics, Burlingame, Calif). MAP-tau was previously validated by immunoblot analysis and siRNA knockdown,4 and stathmin was validated in this laboratory by immunoblot analysis and siRNA knockdown (Fig. 1A,B). Serial sections of the index array TMA were stained alongside each Yale University cohort slide to confirm assay reproducibility. Normal breast epithelium in the Yale University cohort TMA served as internal positive controls, whereas omission of the primary antibody served as the negative control for each immunostaining event. Immunofluorescence staining was performed as described.36

TMA Image Capture and Analysis

Details of the AQUA method of quantitative immunofluorescence has been previously described.36 The TMA cohorts were captured and analyzed using version 1.6 of the AQUA software (HistoRx, Branford, Conn) on the PM2000 platform. Specific parameters related to the TMA data collection are found in Giltnane et al.36

Statistical Analysis

Average values for MAP-tau and stathmin AQUA scores were calculated from 2-fold redundant samples and treated as independent continuous variables. Survival curves for both cohorts were constructed using Kaplan-Meier methods, and the Cox-Mantel log-rank test was used to calculate the association between expression and survival. Cox proportional hazards regression analysis was used to determine which independent factors significantly affected overall survival (OS) in the Yale University cohort. All P values were based on 2-sided testing and differences were considered significant at P < .05. Statistical analysis was performed using JMP Statistical Discovery Software, version 7.0.1 (SAS Institute, Cary, NC).

RESULTS

Stathmin Antibody Validation

Stathmin expression was evaluated in 10 breast cancer cell lines and in a colorectal cancer cell line by western blot and TMA. Western blots showed bands at the appropriate migration distances compared to molecular weight standards (Fig. 1A). AQUA scores were collected for an overlapping series of cells lines that showed a range of expression with lowest expression observed in SKBR3 cells (AQUA score = 63) and highest in UACC 812 cells (AQUA score = 1903). In addition, stathmin siRNA knockdown was performed in a BT-20 cell line, indicating decreased stathmin expression after 24 hours (Fig. 1B) and providing evidence of antibody specificity. Finally, 2 different lots of stathmin antibody were purchased and tested on breast TMAs cored from the same tissue block to confirm reproducibility of the antibody.

Stathmin Expression Pattern in Normal Breast Tissue and the Yale University Cohort

Stathmin was measured in normal epithelial ducts and lobules using a normal breast TMA (n = 110) with 2-fold redundancy to determine the level of expression in normal breast tissue. Consistent cytoplasmic localization was observed, and stathmin average expression scores in normal breast tissue showed mean and median AQUA scores of 25 and 18, respectively, and a score range of 13 to 127. A frequency distribution of these scores is shown (Fig. 1C). Normal tissue TMAs and an index TMA were also used to determine the threshold of signal to noise. We found that AQUA scores below 18 represent nonspecific or background signal (images not shown). The threshold for detection was determined by measuring a control slide in which the primary antibody was omitted. All subsequent AQUA scores were normalized to this AQUA score range using an index TMA.

Next, we examined stathmin expression within the Yale University cohort (n = 651), a primary breast cancer cohort previously used to study many other biomarkers.33, 36, 37 Stathmin showed cytoplasmic localization similar to our observations in normal breast tissue. A stathmin AQUA score was computed for each case by averaging observations from 2 TMA spots and normalizing them, using the index TMA. The frequency distribution of stathmin expression scores in the Yale University cohort is shown along with the mean and median AQUA scores of 62 and 30, respectively, with a score range of 4 to 1092 (Fig. 1C). A total of 474 cases had sufficient tumor tissue for analysis. The mean stathmin expression score in normal breast tissue was used in all analyses to stratify patients as high expressers (AQUA score ≥ 25) versus low expressers (AQUA < 25) with 41.8% classified as low stathmin expressers compared with 58.2% classified as high expressers in the Yale University cohort (Table 1).

Table 1. Univariate Analysis of Tumor and Clinical Risk Factors for Overall Survival in the Yale University Cohort
  Univariate
VariableNo. of Patients (%) (n = 651)HR95% CIPa
  • CI indicates confidence interval; ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; HR, hazard ratio; MAP, microtubule-associated protein; PR, progesterone receptor.

  • a

    P is given for Cox univariate analysis, statistically significant P values (P < .05) are in bold, trending P values are in italics

Age at diagnosis645 (99.1)1.0080.996-1.020.178
 Unknown/missing6 (0.9)   
Menopausal status    
 Premenopausal196 (30.1)1.000 .021
 Postmenopausal449 (69.0)1.4341.055-1.950 
 Unknown/missing6 (0.9)   
Tumor size (cm)    
 ≤ 2215 (33.0)1.0000.204-0.446<.0001
 > 2384 (59.0)2.1581.555-2.996 
 Other52 (8.0)   
Nodal status    
 Negative for node metastasis327 (50.2)1.000 <.0001
 Positive for node metastasis320 (49.2)2.3381.761-3.104 
 Unknown/missing4 (0.6)   
Nuclear grade    
 Small/uniform nuclei113 (17.4)1.000 .024
 Intermediate nuclei315 (48.4)1.1120.705-1.756 
 Large nuclei170 (26.1)1.8531.158-2.967 
 Other53 (8.1)   
ER status    
 ER-positive326 (50.1)1.000 .015
 ER-negative289 (44.4)1.4021.257-2.288 
 Other36 (5.5)   
PR status    
 PR-positive302 (46.4)1.000 .013
 PR-negative294 (45.2)1.4211.075-1.877 
 Other55 (8.4)   
HER2 status    
 HER2-positive109 (16.7)1.000 .606
 HER2-negative495 (76.0)0.9130.646-1.290 
 Other47 (7.2)   
MAP-tau expression    
 MAP-tau low expression376 (57.8)1.000 .0042
 MAP-tau high expression104 (16.0)0.6910.489-0.974 
 Unknown/missing171 (26.3)   
Stathmin expression    
 Stathmin low expression198 (30.4)1.000 .0061
 Stathmin high expression276 (42.4)1.4831.119-1.966 
 Unknown/missing177 (27.2)   
Ratio for MAP-tau to stathmin    
 Low MAP-tau:high stathmin237 (36.4)1.000 .0053
 High MAP-tau:low stathmin237 (36.4)0.6790.517-0.891 
 Unknown/missing177 (27.2)   

Premenopausal status, high nuclear grade, and ER-negative status correlated most frequently with high stathmin expression compared to low expression (34.4% vs 22.6%, P = .005; 34.6% vs 22.12%, P = .0004; and 50.6% vs 38.9%, P = .0136, respectively). Stathmin expression did not correlate with tumor size, nodal status, or PR or human epidermal growth factor receptor 2 (HER2) status (data not shown).

Stathmin Prognostic Value in the Yale University Cohort

The expression status of stathmin was evaluated for association with OS. Using Kaplan-Meier analysis, patients with low stathmin expression (n = 195) showed improved survival compared with those who had high expression (n = 273; 66.6% vs 53.3%, respectively; log-rank, P = .003; Fig. 2A). When stratified by ER status, stathmin retained prognostic value, with ER-positive/low stathmin expressers (n = 114) showing improved survival compared to ER-negative/high stathmin expressers (n = 136; 70.9% vs 48%; log-rank, P = .012; data not shown). Similarly, patients stratified by HER2 status indicated improved survival, with HER2-negative/low stathmin expression (n = 156) compared to HER2-positive/high stathmin expression (n = 60) (66.4% vs 53.0% respectively; log-rank, P = .006).

Figure 2.

Kaplan-Meier survival analysis is shown in the Yale University cohort. (A) Shown is the 10-year survival for high stathmin expression (n = 273) versus low expression for all invasive breast carcinoma patients in the Yale University cohort. An AQUA score of 25 was used as the cutpoint based on the mean expression score from normal breast tissue. (B) Kaplan-Meier survival is shown for MAP-tau expression in the Yale University cohort with cohort division using the cutpoint based on MAP-tau expression in normal breast tissue.38 (C) Survival analysis is shown for the ratio of MAP-tau to stathmin using the median ratio score as the cutpoint to differentiate high MAP-tau/stathmin expression ratio. (D) Frequency distribution is shown for AQUA ratio scores in the Yale University cohort; dotted arrow shows stratification of the cohort using the median score.

In univariate analysis, high stathmin expression, postmenopausal status, nodal metastasis, and ER- and PR-negative status were associated with worse OS (hazard ratio [HR] = 1.483, HR = 1.434, HR = 2.338, HR = 1.402, HR = 1.421; P = .0006, P < .001, P < .015, and P < .013, respectively), whereas small tumor size and low nuclear grade were associated with improved OS (HR = 0.493 and HR = 0.614; P < .001 and P < .024, respectively; Table 1). Using the Cox proportional hazards, multivariate analysis indicates that high stathmin expression has independent prognostic value with an HR of 1.566 (95% CI = 1.091-2.248, P = .015; Table 2). Large tumor size, nodal metastasis, and ER-negative status were also associated with worse OS.

Table 2. Multivariate Analysis of Tumor and Clinical Risk Factors for Overall Survival in the Yale University Cohort
 Multivariate
 MAP-tauStathminRatio of MAP-tau to Stathmin
VariableNo. of Patients (%) (n = 651)HR95% CIP*HR95% CIP*HR95% CIP*
  1. aP is given for Cox multivariate analysis, statistically significant P values (P < .05) are in bold, and trending P values are in italics.

  2. CI indicates confidence interval; ER, estrogen receptor; HER2, human epidermal growth factor receptor 2; HR, hazard ratio; MAP, microtubule-associated protein; PR, progesterone receptor.

Age at diagnosis645 (99.1)1.0040.981-1.026.7180.9980.975-1.021.8440.9990.977-1.022.952
 Unknown6 (0.9)         
Menopausal status          
 Premenopausal196 (30.1)1.000 .3391.000 .2051.000 .357
 Postmenopausal449 (69.0)1.1580.857-1.569 1.4890.804-2.759 1.3360.721-2.474 
 Other6 (0.9)         
Tumor size (cm)          
 ≤ 2215 (33.0)1.000 <.00011.000 <.00011.000 <.0001
 > 2384 (69.0)2.0691.459-2.934 2.0171.421-2.862 2.0421.440-2.896 
 Unknown52 (8.0)         
Nodal status          
 Negative for node metastasis327 (50.2)1.000 <.00011.000 <.00011.000 <.0001
 Positive for node metastasis320 (49.2)1.5781.307-1.922 2.6271.791-3.854 2.7161.849-3.989 
 Unknown4 (0.6)         
Nuclear grade          
 Small/uniform nuclei113 (17.4)1.000 .4211.000 .19491.000 .296
 Intermediate nuclei315 (48.4)1.0150.624-1.725 0.6760.374-1.222 0.8510.460-1.575 
 Large nuclei170 (26.1)1.2720.868-1.853 1.1860.738-1.906 1.2510.760-1.940 
 Unknown53 (8.1)         
ER status          
 ER-positive326 (50.1)1.000 .3081.000 .0311.000 .101
 ER-negative289 (44.4)0.9030.740-1.097 1.5211.037-2.231 1.3800.939-2.029 
 Unknown36 (5.5)         
PR status          
 PR-positive302 (46.4)1.000 .4331.000 .2021.000 .395
 PR-negative294 (45.2)0.9290.769-1.114 1.2530.886-1.771 1.1650.819-1.657 
 Unknown55 (8.4)         
HER2 status          
 HER2-positive109 (16.7)1.000 .0201.000 .0671.000 .102
 HER2-negative495 (76.0)1.2921.041-1.586 0.6750.443-1.029 0.7040.461-1.073 
 Unknown47 (7.2)         
MAP-tau expression          
 MAP-tau low expression376 (57.8)1.000 .018
 MAP-tau high expression104 (16.0)0.7650.598-0.957 
 Other171 (26.3)         
Stathmin expression          
 Stathmin low expression198 (30.4)1.000 .0150
 Stathmin high expression276 (42.4)1.5661.091-2.248 
 Other177 (27.2)         
Ratio for MAP-tau to stathmin          
 Low MAP-tau:high stathmin237 (36.4)1.000 .0081
 High MAP-tau:low stathmin237 (36.4)0.6090.422-0.879 
 Other177 (27.2)         

Prognostic Value for the Ratio of MAP-Tau to Stathmin in the Yale University Cohort

We have shown that high MAP-tau is associated with favorable outcome (Fig. 2B). Because MAP-tau and stathmin play opposite but complementary roles in tubulin stabilization, we assessed the ratio of MAP-tau to stathmin expression levels. The frequency distribution of ratio scores for MAP-tau to stathmin expression in our cohort showed mean and median AQUA scores of 13 and 7, respectively, with a ratio score range of 0.098 to 184 (Fig. 2D). The median ratio score of 7 was used as the cutpoint to differentiate high MAP-tau/stathmin expressers from low MAP-tau/stathmin expressers (Table 1). Low MAP-tau/stathmin expression ratio correlated with premenopausal status, high nuclear grade, ER- and PR-negative status, and HER2-positive status.

Using Kaplan-Meier survival analysis, patients with high MAP-tau/stathmin expression (n = 231) showed improved survival compared with those who had low MAP-tau/stathmin expression (n = 237; 65.4% vs 52.5%, respectively; log-rank, P = .0009; Fig. 2C). When patients were stratified by ER status, the MAP-tau ratio has no prognostic value in ER-positive patients. However, in ER-negative patients, high MAP-tau/stathmin expression showed improved survival compared to ER-negative and low MAP-tau/stathmin expression (n = 139; 66.1% vs 45.5%; log-rank, P = .009). When patients were stratified by HER2 status, improved survival with high MAP-tau:low stathmin expression was seen in both HER2-positive and HER2-negative subgroups.

In univariate analysis, a high MAP-tau/stathmin expression ratio was associated with improved OS (HR = 0.679; 95% CI = 0.517-0.891; P = .0053; Table 1). The ratio was also significant by multivariate analysis, where patients with high MAP-tau/stathmin expression had an HR of 0.609 (95% CI = 0.422-0.879; P = .008; Table 2).

As a destabilizing protein, we hypothesized that stathmin would have an inverse relationship to MAP-tau. Figure 3A shows that is not seen in normal breast tissue. However, in breast tumors (Fig. 3B), there are many cases with high stathmin and low MAP-tau and vice versa, and they define a subset of cases with the hypothesized inverse relationship. However, there are also many cases that are low for both proteins, so no inverse relationship is seen. Because these cases predominate, no statistically significant inverse relationship is seen when assessing the whole cohort.

Figure 3.

Correlation between stathmin expression and MAP-tau expression is shown in normal versus Yale University cohort breast tissue. (A) Correlation between stathmin and MAP-tau expression AQUA scores in normal breast tissue shows no correlation. (B) Correlation between stathmin and MAP-tau expression AQUA scores in the Yale University cohort suggests an inverse correlation.

DISCUSSION

Stathmin expression has been evaluated extensively in a variety of cell lines and has been correlated with prognosis, through use of messenger RNA levels.2, 7, 29, 31 However, only a single small study using traditional IHC has evaluated stathmin protein expression, with stathmin expression examined as a surrogate marker for a PTEN gene expression signature.32 In that study, Saal and colleagues found stathmin IHC staining scores were significantly higher in PTEN-negative tumors than in PTEN-positive tumors (P = .005), indicating that loss of the tumor suppressor PTEN, and subsequent activation of the oncogenic phosphoinositol 3-kinase pathway, was associated with increased stathmin expression. In addition, high-stathmin-expressing patients experienced significantly worse distant disease-free survival than low-stathmin-expressing patients. This study confirms these results using a larger cohort.

Because entry and exit from mitosis requires the coordinated activity of both microtubule stabilizers and destabilizers, we hypothesized that a 2-marker model would provide improved prognostic information for patients with breast cancer that is not currently available through single-marker evaluation of either MAP-tau or stathmin alone. Previous evaluation of MAP-tau in the Yale University cohort provided an opportunity to generate a combined tissue biomarker containing information from both a microtubule stabilizer and destabilizer. This analysis showed that the high MAP-tau/stathmin ratio is prognostic for improved survival in patients with breast cancer. When the evaluation of MAP-tau is compared to the expression levels of stathmin, an inverse relationship is seen in a subset of the cases. When combined as a ratio, the combined variable is more prognostic than either variable alone, although the improvement is not statistically significantly better.

A key limitation of this work is the lack of a taxane-treated cohort in which to assess the predictive power of either stathmin or the MAP-tau/stathmin ratio for response to taxane therapy. Although we were able to access a taxane trial that showed no predictive value for MAP-tau,38 the tissue from that cohort is exhausted. Other cohorts are being sought to investigate stathmin and other variables related to microtubule stability.

Because of varying response rates to taxanes and significant adverse effects in patients with breast cancer, a method to predict response to these chemotherapeutic agents is desirable. Currently, no diagnostic test is recommended to differentiate patients who would benefit from taxane therapy from those who could avoid its potential cytotoxic effects. Recently, beta III tubulin was assessed as a candidate companion diagnostic test, but has not proven valuable after disappointing clinical trial results.39–41 However, a new and promising marker, TLE3 (transducin-like enhancer of split 3), is now being assessed for usefulness in selecting patients who respond to taxanes. Data in single-institution studies in both lung and ovarian cancer show that TLE3 levels correlate with response to therapy.42, 43

In summary, the present study uses a quantitative method to assess stathmin expression and finds a significant prognostic relationship, with high stathmin expression associated with worse outcome in patients with breast cancer. Furthermore, combining this work with previous quantitative assessments of MAP-tau allowed the construction of a novel tissue biomarker reflecting opposite but complementary cellular roles for MAP-tau and stathmin. In our analysis, the ratio of MAP-tau to stathmin expression is associated with improved OS and showed greater prognostic magnitude than either marker examined individually. Future efforts to test this ratio for predictive value in anti-microtubule–treated patients are indicated.

FUNDING SOURCES

This work was supported by Breast Cancer Research Program grant #W81XWH-06-1-0746 from the US Army Medical Research and Materiel Command to M.T. Baquero. A.M. Molinaro was supported by Clinical and Translational Science Award grant UL1 RR024139 from the National Center for Research Resources.

CONFLICT OF INTEREST DISCLOSURE

D.L. Rimm is a stockholder in and consultant to HistoRx, the exclusive licensee to the Yale-owned AQUA technology. A.M. Molinaro has served as a paid consultant to HistoRx.

Ancillary