Prostate carcinoma is a dynamic process in which different phases in the natural history of the disease may be characterized by unique biologic mechanisms.1 The pathways that underlie tumor pathogenesis, growth, and resistance to treatment in each clinical state increasingly are understood, and novel biologic agents that target these pathways are now available for clinical testing. Trastuzumab (Herceptin™; Genentech, San Francisco, CA) is a humanized monoclonal antibody that targets the extracellular domain of HER-2, a glycoprotein that shares homology with the epidermal growth factor receptor and that has been implicated in malignant transformation and tumorigenesis.2–8 Targeting HER-2 using trastuzumab in combination with chemotherapy confers a survival benefit for patients with metastatic breast carcinoma that overexpresses HER-2,9 raising the possibility of therapeutic benefits for other types of malignancies that similarly may express HER-2.10, 11
Profiling of HER-2 status has shown a range of findings in prostate carcinoma. The reported range of HER-2 positivity in androgen dependent (AD) samples varies widely, depending on tissue selection and preparation; the technique used to determine positivity; the antibody used for immunohistochemical analysis; and the distribution, degree, and intensity of staining needed to characterize a specimen.12–14 Alterations in HER-2 expression as a tumor progresses from localized to metastatic disease and from androgen dependence to androgen independence has yet to be fully established; preliminary data suggest that expression is not uniform across disease states.15–18
Preclinical data in prostate carcinoma suggest a role for targeting HER-2 in a variety of clinical states. Craft et al. reported that HER-2 could effect ligand independent activation of the androgen receptor,19 suggesting a role for trastuzumab in patients with androgen independent (AI) disease, whereas animal models suggest that trastuzumab as a single agent is effective in patients with AD disease.20
This study explored the feasibility of molecular profiling in prostate carcinoma to determine the correlation between hormonal sensitivity, HER-2 expression, and the antitumor effects of trastuzumab alone and with paclitaxel. Patients were treated with trastuzumab alone until they experienced disease progression, when paclitaxel was added. Paclitaxel was selected on the basis of its activity in prostate carcinoma alone21 and in combination with other drugs22–25 as well as for its demonstrated synergistic effects with trastuzumab.26
MATERIALS AND METHODS
Pathology Staining and Eligibility
To participate in this trial, patients were required to have progressive metastatic prostate carcinoma. Tissue from registered patients was tested for HER-2 expression. Overexpression could be documented either on the original diagnostic biopsy or on a biopsy of recurrent or metastatic disease. All patients' tumors were stained for HER-2 using the Herceptest™ immunohistochemistry kit (Dako, Carpinteria, CA). A single reference pathologist interpreted all biopsies. Tumors in which HER-2 was assessed at < 2+ (moderate, circumferential membrane staining readily visible at ×10 magnification and observed in > 10% of tumor cells) with the standard Dako kit were considered as not overexpressing the protein. If multiple tissue samples were available and these conflicted in terms of HER-2 status, then assignment to an arm was made on the basis of the HER-2 positive result.
Patients were assigned to one of four arms: AI HER-2 positive, AI HER-2 negative, AD HER-2 positive, and AD HER-2 negative. To be eligible for the AI arm, patients must have progressed on hormonal treatment despite antiandrogen withdrawal, when applicable. AD patients were eligible if they received neoadjuvant hormonal therapy prior to primary radiation therapy or intermittent hormonal therapy, as long as the testosterone level exceeded 50 ng/mL prior to registration.
Treatment and Evaluations
All patients were treated with an initial loading dose of 4 mg/kg of trastuzumab administered over 90 minutes. Subsequently, weekly doses of 2 mg/kg were given over 30 minutes in 12-week cycles. At progression, weekly paclitaxel at 100 mg/m2 given over 1 hour was added to the regimen. Patients received a maximum of 3 12-week cycles of therapy. Patients with AD disease remained off hormones while on the study.
Patients underwent regular complete blood counts, routine blood chemistries, assays of serum concentrations of prostate specific antigen (PSA), and physical examinations. Toxicities were assessed according to the National Cancer Institute Common Toxicity Criteria (version 2.0). At the end of each 12-week cycle, the patient's index lesions were reimaged by the modality that was performed at baseline.
We anticipated that trastuzumab would have a cytostatic effect rather than a tumoricidal effect. Therefore, the primary end point was the proportion of patients who showed no rise in PSA at 12 weeks (responders) compared with patients who showed a rise in PSA (nonresponders), with no other signs of disease progression either clinically or by imaging studies. This end point was supported by a mathematic model based on a data set using 254 patients that we used previously to show that a decline in PSA > 50% over a 12-week period correlated with a survival benefit.27 Using the same 254 patients, we found that a survival benefit was conferred even when a standard of no rise in PSA was applied. This was validated against an independent data set using the same methodology cited above (our unpublished results).
Each arm accrued and was analyzed independently. The target accrual was 14 patients per arm; if no responses were observed, then accrual to that arm would stop. If 1 response was observed, then that arm would expand to 25 patients. This decision ensured that if the probability of response in the population was ≥ 0.20, there would be < 5% probability of stopping the trial after the first stage. In addition, the cohort with AI HER-2 negative disease was analyzed after accrual and treatment was complete to determine whether treatment had an impact on the slope of the log of the PSA. A 95% bootstrap-t confidence interval was constructed to delineate the difference in the log of pretreatment and post-treatment (with monotherapy) PSA slopes using 5000 bootstrap samples.28
This trial was approved by the Memorial Sloan-Kettering Cancer Center Institutional Review Board on December 1, 1998, and began accruing patients on January 20, 1999. The last patient began treatment on November 10, 1999. Informed consent was obtained from all patients.
The numbers of screened patients and treated patients are described in Figure 1. One hundred thirty patients were screened. The median age was 68 years (range, 47–84 years). Seventeen patients had AD tumors, and 112 patients had AI tumors; we had insufficient information on one patient to determine the hormonal sensitivity of his tumor. Of the screened patients, we could not acquire tumor-containing tissue from 24 patients, because a biopsy failed to produce tumor and/or archived tissue could not be obtained. One hundred seventeen tissue samples were obtained from the remaining 106 patients for HER-2 testing.
Because patients with HER-2 positive disease were not identified readily (and not all patients with HER-2 positive tumors met the eligibility criteria otherwise), a total of 23 patients received treatment. The demographics of these patients are described in Table 1. Of the 20 patients with AI disease, 6 patients received 1 prior hormonal therapy, 14 patients received 2 or more prior hormonal regimens, and 9 patients received prior chemotherapy. All three patients with AD disease received prior neoadjuvant hormones. Table 2 describes the distribution of treated patients to the four treatment arms. Six patients were in the AI HER-2 positive cohort, and 14 patients were in the AI HER-2 negative arm. The remaining three patients in the study had AD disease, all with HER-2 negative tumors.
|Median Age (range)||66 years (49–84 yrs)|
|Median Karnofsky performance status (range)||90 (70–90)|
|Baseline median biochemical parameters (range)|
|Hemoglobin (g/dl)||12.8 (10.5–16.4)|
|LDH (U/L)||158 (63–316)|
|PSA (ng/mL)||66.8 (6.49–1647.29)|
|Androgen independent (n = 20 patients)|
|Androgen dependent (n = 3 patients)|
Treatment and Efficacy
An overview of treatments and outcomes for all patients is described in Table 2. Of the 14 patients with AI HER-2 negative tumors, 13 patients developed disease progression within the first 12-week cycle of single-agent therapy with trastuzumab. One patient was given a second cycle of monotherapy, but his PSA level began rising in the 13th week on study (prior to new lesions appearing on bone scan); therefore, he was considered a nonresponder. To determine whether single-agent therapy had an impact on the slope of the PSA rise in these patients, we constructed a 95% bootstrap-t confidence interval for the difference in log PSA slopes between pretreatment and post-treatment with trastuzumab. Using 5000 bootstrap samples, the confidence interval had a lower bound of −0.0465 and an upper bound of 0.0516, signifying no difference between pretreatment log PSA slope and log PSA slope after treatment with trastuzumab.
Ten patients with AI HER-2 negative disease received trastuzumab and paclitaxel, and 9 patients were evaluable for response. Of these 9 patients who received both drugs, 5 patients (56%) experienced disease progression at or before the 12-week evaluation period, 3 patients (33%) had stable disease, and 1 patient (11%) patient had a partial response.
Six patients with AI HER-2 positive disease were treated. Of these, all progressed on single-agent trastuzumab at or before the 12-week landmark. Four patients received combination therapy: Two patients had a partial response by PSA, and two patients had disease progression.
Three patients with AD disease were treated, all of whom had HER-2 negative tumors. These patients experienced disease progression on monotherapy within the initial 12 weeks of treatment. Two patients received combined treatment, during which time they had stable disease.
Treatment with trastuzumab alone and with paclitaxel was well tolerated, as shown in Table 3. For patients who received monotherapy, the only Grade 4 toxicity was an acute myocardial infarction in one patient. It was felt that this was be unrelated to drug, because it was an acute ischemic event without evidence of the expected cardiomyopathy associated with trastuzumab treatment. No Grade 3 toxicities were drug related.
|Fatigue/malaise||13 (57)||3 (13)||1 (4)||0||9 (56)||6 (38)||0||0|
|Left ventricular function||2 (9)||1 (4)||0||0||1 (6)||0||0||0|
|Dyspnea||0||5 (22)||0||0||4 (25)||3 (19)||1 (6)||0|
|Nausea/emesis||7 (30)||0||0||0||7 (44)||0||1 (6)||0|
|Paresthesias||7 (30)||0||0||0||6 (38)||4 (25)||2 (13)||0|
|Hemoglobin||10 (43)||4 (17)||0||0||7 (44)||5 (31)||1 (6)||0|
|Neutrophils||3 (13)||1 (4)||0||0||6 (38)||2 (13)||1 (6)||1 (6)|
|Platelets||4 (17)||1 (4)||0||0||3 (19)||1 (6)||0||0|
With the addition of paclitaxel, all patients were hyperglycemic (Grade 3 in five patients), probably as a result of steroid premedication. No toxicities beyond those ordinarily seen with weekly paclitaxel were seen. In addition, two patients on combination therapy had lower extremity deep venous thromboses. Whether this was due to underlying disease or to therapy is unclear.
One hundred seventeen biopsy specimens from 106 patients were analyzed and are described in Table 4. These biopsies were categorized by androgen sensitivity at the time of the biopsy, not at the time of treatment. In total, 90% of specimens were HER-2 negative, and 10% of specimens were HER-2 positive. Ninety-seven biopsies were from the prostate, and 20 biopsies were from metastatic sites. Of the prostate biopsies, 84 were from patients with AD disease, and 13 were from patients with AI disease. None of the AI local prostate biopsies overexpressed HER-2. Six of 84 AD prostate biopsies (7%) overexpressed HER-2. Two of the prostate biopsies from the same patient varied in HER-2 expression: The AD sample was HER-2 positive, and the AI sample was HER-2 negative. Metastatic tissue for this patient was unobtainable.
|No. of samples||13||84||12||8||117|
|Her-2+||0 (0%)||6 (7%)||5 (42)||1 (12)||12 (10% of all specimens were HER-2 positive)|
|Her-2−||13 (100%)||78 (93%)||7 (58)||7 (88)||105 (90% of all specimens were HER-2 negative)|
Of 20 metastatic lesions, 12 lesions were from patients with AI disease, 5 of which (42%) were HER-2 positive, and 7 of which (58%) were HER-2 negative. Eight of 20 metastatic lesions were from AD patients. Of these, one lesion (12%) was HER-2 positive.
Overall, there were nine matched pairs in which both the AD primary prostate specimen and the AI metastatic tissue were available for HER-2 testing. In three samples, the AI metastases were HER-2 positive, and the corresponding AD prostate samples were HER-2 negative. Six matched pairs were concordantly HER-2 negative.
This trial was designed to examine the feasibility of molecular profiling in patients with prostate carcinoma to determine the correlation between hormonal sensitivity, HER-2 expression, and the antitumor effects of trastuzumab alone and with paclitaxel. Although only patients in the HER-2 negative AI metastatic arm accrued to completion, we are reporting the pathology and the incomplete clinical data from the other arms, because these results argue against seeking additional HER-2 positive patients until new methods of acquiring metastatic tissue to screen for HER-2 expression are developed. Given the plethora of studies using biologic agents, our experience suggests that screening for the targeted pathway must occur on tissue that represents the disease at the time of treatment, not at diagnosis.
Our clinical results reveal that trastuzumab is not effective for patients with AI HER-2 negative disease. Although stable disease or partial responses were seen in 8 of 15 patients who received trastuzumab and paclitaxel, the value of adding trastuzumab to the chemotherapy regimen is uncertain, because paclitaxel alone can induce response proportions in the 50% range.21
Our pathology results prospectively affirm what retrospective analyses have suggested previously: that HER-2 expression in prostate carcinoma varies by clinical state. We screened 130 patients but identified only 6 eligible patients with HER-2 positive disease. HER-2 overexpression was found in greater proportion in AI metastases (5 of 12 samples; 42%) compared with prostatic tissue or even AD metastatic tissue, although the number of specimens was small. In three of nine matched pairs, the AD prostate biopsy did not overexpress HER-2 but the AI metastatic sample did, providing evidence that variation in HER-2 expression can be documented even within individual patients as their disease progresses. These results confirm the findings of Signoretti et al, who found that HER-2 is overexpressed in 25% of untreated prostate carcinoma specimens, in 59% of specimens treated with hormones before surgery, and in 78% of patients with AI disease.18 In addition, our group found that HER-2 was overexpressed in 20% of untreated primary prostate specimens, in 67% of hormonally treated prostates, and in 80% of metastatic lesions.29 In aggregate, these data suggest that accurate HER-2 profiling of patients with metastatic disease requires testing tissue at all points in the history of the natural and treated history of the disease, with particular focus on acquiring metastatic tissue. The low rate of HER-2 expression found in this trial is explained in part by having tested primarily AD primary prostate specimens, although we found that AI prostate specimens had little HER-2 overexpression as well.
Acquiring metastatic tissue from patients with prostate carcinoma is problematic. In our group's experience, blind bone marrow biopsies yield tumor in only 30% of patients. Even if HER-2 overexpression was present in 40% of those patients, identifying 14 patients with HER-2 positive disease would require over 100 biopsies (over 200 biopsies if we had a single response and needed to accrue 25 patients). Another option would be to screen only patients with soft tissue disease, although such patients are in the minority, and tumors in lymph nodes may differ biologically from those in bone. One proposed solution to the difficulties of obtaining metastatic tissue for screening is to use shed serum HER-2 antigen.30, 31 We performed shed HER-2 assays on serum samples as a post-hoc analysis, but we did not have enough biopsy data on metastatic samples to make meaningful correlations between the two methods. Given these logistic issues, we opted not to pursue further efforts to accrue further patients with HER-2 positive disease.
Even if metastatic tissue is available for screening, the most accurate means of determining the HER-2 status is controversial, nor is it established which method of HER-2 testing, if any, predicts a therapeutic response in patients with prostate carcinoma.32 Studies comparing HER-2 testing of prostate carcinoma specimens using fluorescence in situ hybridization (FISH) with immunohistochemistry found that FISH was more sensitive.13, 14 Others have found no gene amplification, even in specimens with protein overexpression, suggesting that FISH is a limited tool for screening.18 The optimal assay or combination of tests for HER-2 testing has yet to be determined. We used a criterion of 2+ or greater expression to define overexpression, identical to that of the pivotal trial for metastatic breast carcinoma, although the Herceptest differs from the clinical trials assay used in the breast carcinoma trial.33
This study underscored the difficulties of conducting clinical trials of biologic agents in patients with prostate carcinoma. Such trials face the four-fold challenge of developing validated assays of active pathways, acquiring tissue, establishing that a pathway is clinically relevant at the time of treatment, and determining meaningful end points. Although HER-2 remains a valid therapeutic target, the data from this trial suggest that trastuzumab is not active as a single agent for the treatment of patients with AI HER-2 negative disease and that further pursuit of patients with HER-2 positive disease for clinical testing is not feasible logistically using traditional immunopathologic methods. Future trials of biologic agents in patients with metastatic prostate carcinoma will require new techniques for assessing biologic pathways in metastatic tissue, and the assays used to test for these pathways require further validation.
HER-2 (also known as ErbB2 or neu) belongs to the ErbB family of 4 type I tyrosine kinase receptors, including epidermal growth factor receptor (EGFR), HER-3, and HER-4, that homo- and heterodimerize to activate distinct programs of proliferation, survival, migration, and angiogenesis.1 In breast cancer, this family also demonstrates cross-talk with the hormone receptors for estrogen (ER) and progesterone (PR), as well as other pathways.2
ErbB2 amplification is an important molecular alteration in breast cancer, and we hypothesized that interactions of HER-2 with other ErbB family members might improve our ability to classify HER-2+ breast cancer for the purposes of prognosis. Although members of the HER family have been measured for prognostic value,3-12 they have never previously been rigorously multiplexed, because nearly all previous studies have been scored by traditional pathologist-based methods.
To measure tumor-specific content, most quantitative protein measurement techniques such as mass spectrometry require microdissection and are optimized with frozen specimens. Automated quantitative analysis (AQUA) measures protein expression levels in situ in formalin-fixed, paraffin-embedded tumor samples and allows discrimination of subcellular compartments.13, 14 The AQUA method has been validated in breast cancer and shown to be comparable to protein levels measured by enzyme-linked immunosorbent assay (ELISA),15, 16 with coefficients of variation <5%. To test the hypothesis that quantitative multiplexed analysis will improve prognostic value, we have assessed the expression of 6 targets (ER, PR, EGFR, HER-2, HER-3, HER-4) in 4 subcellular compartments, using AQUA in an archival tissue microarray (TMA) collection of invasive breast carcinoma.
MATERIALS AND METHODS
A TMA containing cores from formalin-fixed, paraffin-embedded cell pellets was used as a control for staining and AQUA analysis. JEG-3, SKOV3, and CHO cells were obtained from the Maihle laboratory at Yale University. A431, HL60, MDA-MB-453, MDA-MB-231, MDA-MB-468, SW-480, SK-BR-3, MCF-7, BT-549, T-47D, MDA-MB-435S, and BT-474 cell lines were purchased from the American Type Culture Collection (Manassas, Va). BAF3 cells were obtained from a laboratory in the Department of Genetics at Yale University, and culture conditions and cell line TMA construction have been published in detail elsewhere.15, 16 Our laboratory protocol for processing cell lines is also available on the Web (http://www.tissuearray.org).
The Yale breast cancer cohort consists of 676 samples of invasive breast carcinoma collected serially from the Yale University Department of Pathology archives from 1961 to 1983 (Table 1). Slides were reviewed for tumor volume, and all samples were included that could be adequately sampled for the study. This cohort contains approximately half lymph-node–positive specimens and half lymph-node–negative specimens. Patient outcome was collected from the medical records and the Connecticut Tumor Registry. TNM classification was applied retrospectively according to guidelines from the AJCC Cancer Staging Manual, Sixth Edition.17 For the 630 patients with outcome data, the mean follow-up time is 12.5 years, and the mean age at diagnosis is 58.1 years. The median follow-up time is 8.8 years, and the median age at diagnosis is 58.0 years. A total of 334 patients were censored at 10 years, and 228 were uncensored at 10 years. Of the 334 censored patients, their median follow-up was 18.9 years, with the minimum at 4.2 months. Complete treatment information was not available for the entire cohort; however, most patients were treated with postsurgical local irradiation. None of the lymph-node–negative patients was given adjuvant systemic therapy.
|Age at diagnosis, y||630||100.00|
Formalin-fixed paraffin-embedded tumor blocks from each patient were used in the construction of the tissue microarrays, with 1 0.6-mm core transferred to a recipient paraffin block. Slides cut from 2 independent constructions were used in this study for each target. A sequential hematoxylin and eosin–stained slide was histologically assessed by a pathologist to ensure adequate tumor sampling. TMA construction was performed with a tissue-arraying instrument (Beecher Instruments, Silver Springs, Md) using a method that was described previously.18 All precut sections were coated in paraffin and stored at room temperature in a nitrogen chamber before staining to prevent loss of antigenicity.19
Slides were stained by a modified indirect immunofluorescence method as described previously.13 Primary antibodies used to define the tumor compartment of each histospot included mouse monoclonal cytokeratin AE1/AE3 (M3515, Dako Corporation, Carpinteria, Calif) or wide-spectrum screening rabbit anticow cytokeratin antibody (Dako Z0622), each at 1:100. Estrogen receptor (Dako clone 1D5) and progesterone receptor (Dako clone PgR636) were each used at 1:50 and incubated for 1 hour at room temperature. Other target antibodies were incubated overnight at 4°C and included EGFR used neat (Dako pharmDx kit clone 2-18C9), HER-2 at 1:8000 (Dako A0485), HER-3 at 1:200 (clone RTJ1, Vector Laboratories, Burlingame, Calif), and HER-4 at 1:400 (sc-283, Santa Cruz Biotechnology, Santa Cruz, Calif). Secondary labeling of targets was performed by signal amplification using horseradish-peroxidase-labeled secondary reagents (species-specific Dako Envision) followed by Cy-5 tyramide incubation. 4′,6-Diamidino-2-phenylindole (DAPI) in an antifading mounting medium was used to stain the nuclear compartment (Prolong Gold, Invitrogen, Eugene, Ore).
Positive and negative controls were included in a specialized “boutique” array stained simultaneously containing 40 cases from a previously described breast carcinoma tissue microarray,16 as well as 15 formalin-fixed, paraffin-embedded cancer cell lines exhibiting variable levels of expression for each marker analyzed. In addition, a breast cancer test slide was stained with each experiment without primary antibody.
AQUA of Tissue Microarrays
A complete and detailed discussion of the AQUA method has been published previously.13, 20 Briefly, monochromatic images of each histospot were acquired on an Olympus AX-51 epifluorescence microscope (Olympus, Melville, NY) using a motor-driven stage and automated custom software, and high-resolution (1024 × 1024 pixel; 0.5 μm) digital images were analyzed using AQUA. A binary image (tumor mask) was created from the cytokeratin image of each histospot, representing areas of tumor epithelium. Histospots were excluded if the tumor mask represented <5% of the total histospot area. DAPI images were used to define the nuclear compartment within each histospot, and the membrane compartment was defined by perimembranous coalescence of cytokeratin immunoreactivity with specific exclusion of the nuclear compartment.
Application of the rapid exponential subtraction algorithm was used to improve subcellular localization; it is an image processing methodology which accounts for compartment overlap because of the thickness of tissue sectioning on glass slides by subtracting out-of-focus from in-focus image data according to a specialized algorithm. Target protein expression was quantified by calculating Cy5 fluorescent signal intensity on a scale of 0-255 within each image pixel. The Cy-5 wavelength is used for target labeling because it is outside the range of tissue autofluorescence. An AQUA score was generated by dividing the sum of target signals within the tumor mask by compartment area. After validation of images to ensure adequate tumor sampling and to exclude any normal epithelium, the AQUA scores were normalized to a 100-point scale and averaged from 2 tumor samples. Although AQUA scores were calculated for each biomarker in 4 subcellular compartments, we restricted survival analysis to the dominant subcellular localization (nuclear: ER, PR; non-nuclear: HER-3; membranous: EGFR, HER-2). In this cohort, HER-4 expression was observed in all 3 compartments, so the total AQUA score in the tumor mask was considered for analysis.
A recent analysis of AQUA for HER-2 measurement showed a strong correlation between AQUA scores, quantitative ELISA protein measurements, and HER-2/neu gene amplification for a standard set of breast cancer cell line controls.15 We repeated both cell line and breast tumor samples used in this study as a reference for HER-2 positivity.
The statistical calculations were performed using JMP Version 5.0 (SAS, Cary, NC). Disease specific survival (DSS) was chosen as the endpoint in the present study. Kaplan-Meier plots were used to illustrate the survival in groups of HER-2+ patients classified by the methods studied, and the log-rank test to test for equality of survival curves. Hazard ratios were estimated using Cox regression. All P values corresponded to 2-sided tests, and values <.05 were considered significant.
Unsupervised hierarchical average-linkage clustering was performed using Cluster and Treeview (Eisen Laboratory, Stanford University, Palo Alto, Calif). Tumors in the Yale cohort that had a value for at least 5 of 6 biomarkers (n = 550) were included in the clustering. AQUA scores were converted to z scores before clustering to normalize between markers.21 For cluster assignment, the distance from dendogram root node was chosen to maximize number of clusters as well as to ensure that each cluster contained at least 5% of the population. No formal statistical test was used to select the number of clusters other than the limitation imposed by the number of subjects in each cluster.
In the AQUA method, cellular compartments and targets are labeled in situ using antibodies conjugated to fluorochrome dyes. Figure 1 shows representative images from immunofluorescent labeling of 6 target biomarkers (ER, PR, EGFR, HER-2, HER-3, HER-4) in the breast cancer samples studied, and each panel shows an enlarged view of the pixel area scored as tumor (tumor mask, Lower Left) and subcellular compartments (Upper Right), as well as Cy5 image for each target (Lower Right). Expression was predominantly nuclear for ER and PR, and membranous for EGFR and HER-2. HER-3 expression was both membranous and cytoplasmic, notably excluded from the nuclear compartment (non-nuclear). HER-4, however, showed 3 distinct patterns of expression: non-nuclear (Fig. 1F), membranous (Fig. 1G), and nuclear (Fig. 1H). Nuclear localization of HER-4 has been described previously, where it is thought to be involved in the transcription of target genes involved in mammary differentiation.22
The YTMA49 cohort is composed of 676 breast cancer cases from the Yale Pathology archives with extensive annotation including long-term follow-up, as described previously (Table 1).16 After standardization by internal controls, AQUA measurements from 2 tumor samples were averaged. Expression data for at least 5 of 6 biomarkers and survival information were available from 550 patients, whereas 126 patients were excluded because of insufficient data.
HER-2 and HER-3 Are Independent Biomarkers of Breast Cancer Survival
As demonstrated in Tables 2 and 3, Cox proportional hazards regression was used to assess the association of each marker with 10-year DSS univariately and in multivariate models. As previously described, high AQUA HER-2 (P = .001) and low AQUA ER (P = .010) and AQUA PR (P = .002) scores were significantly associated with decreased survival.23 In contrast, using ordinal (0-3+) immunohistochemistry (IHC) scores to stratify survival in the same way, low ER (P = .007) and PR (P = .010) scores are associated with decreased survival, but HER-2 expression is not (P = .11). High AQUA EGFR scores trended toward association with decreased survival, but were only of borderline significance (P = .065). In addition, AQUA HER-3 scores were inversely associated with survival (P = .003).
|Age at diagnosis||0.998||0.988-1.009||0.087||.768|
|T2 vs T1||2.080||1.519-2.865|
|T3 vs T2||1.593||1.055-2.347|
|T4 vs T3||0.844||0.471-1.469|
|pN1 vs pN0||2.262||1.623-3.146|
|pN2 vs pN1||1.273||0.872-1.843|
|pN3 vs pN2||1.314||0.855-2.007|
|M1 vs M0||4.646||2.837-7.185|
|2 vs 1||1.214||0.834-1.813|
|3 vs 2||1.626||1.211-2.175|
|A. Individual Models; Each AQUA biomarker+(pT, pN)|
|B. All ErbB Biomarkers+(pT, pN)|
|Cluster group (all vs ErbB-low VI)||18.144||.003|
|T2 vs T1||2.072||1.474-2.931|
|T3 vs T2||1.199||0.754-1.857|
|T4 vs T3||1.025||0.548-1.868|
|pN1 vs pN0||1.743||1.199-2.529|
|pN2 vs pN1||1.264||0.819-1.926|
|pN3 vs pN2||1.249||0.763-2.036|
|M1 vs M0||3.447||1.801-6.102|
The prognostic significance of HER-3 was further explored using the X-tile software program19 to define optimal population cut points in a training set of half the patients in the cohort with both HER-3 expression data and outcome information (n = 260) and validated by Kaplan Meier analysis in the remaining half (n = 261). In the validation set half of the cohort, grouping by HER-3 expression with a cut point of AQUA >25, we find high levels of HER-3 associated with 53% 10-year survival, compared with 69% in the low HER-3 group (log-rank P = .0096).
Next, we constructed a multivariate Cox model including pathological tumor (pT) and nodal (pN) stage with AQUA scores from each of the biomarkers tested and observed that 5 of the 6 biomarkers (ER, PR, EGFR, HER-2, HER-3) were independently correlated with patient outcome when assayed using AQUA. Whereas AQUA ER and AQUA PR were associated with more favorable prognosis, AQUA EGFR, AQUA HER-2, and AQUA HER-3 were associated with decreased survival (Table 3A). When we included AQUA expression scores from all ErbB receptors in a multivariate model including pT and pN , both AQUA HER-2 and AQUA HER-3 remained independent prognostic factors (Table 3B).
ErbB Family Coexpression Is Associated With Prognosis
We used unsupervised average linkage hierarchical clustering to examine the relative coexpression of the AQUA targets measured (Fig. 2). Before clustering, data were normalized for variance between experiments by z score transformation (AQUAz). Six distinct clusters were observed, labeled Cluster I-VI in Figure 2 and colored red (high) to green (low) by the distance from the mean for each target. Clusters I and II were notable for high expression of ER and PR and separated by higher levels of HER-3 and HER-4 in Cluster II. Cluster III was enriched for HER-2 and HER-3 expression and had low levels of hormone receptor expression, whereas high HER-3 and EGFR expression was found in Cluster IV. HER-3 and HER-4 expression was enriched in Cluster V. The largest cluster (VI) included some cases with high expression of EGFR or HER-2, but it had relatively low levels of all targets. Of note, high expression of both EGFR and HER-2 was rarely observed.
We calculated 5- and 10-year DSS rates in the cluster groups. Despite differential ErbB family expression, the high hormone-receptor groups had comparable survival rates (ER-I, 79.8% 5-year and 58.3% 10-year vs PR-II, 77.7% 5-year and 60.4% 10-year). The HER-2/HER-3 Cluster III had the lowest DSS rate (5-year, 44.9%; 10-year, 39.4%). Only 2 of 3 of HER-2–positive patients by conventional IHC (HercepTest 3+) are included in this group, with the majority of remaining HER-2+ patients (20%, 12 of 60) found in Cluster VI. Low survival rates were also observed in the HER-3/EGFR Cluster IV (5-year DSS, 56.2%; 10-year, 42.0%). This group may define the so-called “triple negative” class9, 24, 25 of breast cancer, because these cases are low for ER, PR, and HER-2. The HER-3/HER-4 Cluster V is associated with relatively good outcome (5-year DSS, 70.0%; 10-year, 51.7%). Survival in the low HER family Cluster VI was similar to that in the hormone-receptor–positive Groups I and II (5-year DSS, 74.4%; 10-year, 61.7%). The association of clustering subgroups with outcome was independent of TNM staging parameters when assessed by a multivariate Cox proportional hazards model (P = .003, Table 3C).
Multiplexing AQUA Scores Improves Classification of HER-2+ Breast Cancer
To compare the multiplexed AQUA method to conventional methods, we have done a Kaplan-Meier analysis (Fig. 3). Traditional IHC on the TMA failed to reach significance in this cohort (IHC 3+, Fig. 3A). The addition of quantitative analysis and the use of a previously determined optimized AQUA cut point score16 resulted in an improvement in the prognostic value that achieves statistical significance (Fig. 3B). However, selection of a class of HER-2+ patients defined by hierarchical clustering (Cluster III, Fig. 3C) defines a smaller subset with substantially worse outcome. The median time from diagnosis to death from breast cancer was 98 months in the group defined by IHC, 55 months in the group defined by AQUA, and only 43 months in the group defined by clustering. In contrast, low HER-2 patients had a median survival of almost 200 months by all 3 classification methods.
In this study, we measured the protein expression of the ErbB family (EGFR, HER-2, HER-3, HER-4) and the hormone receptors ER and PR using the AQUA method in a large retrospective cohort of breast cancer patients and assessed target coexpression and association with breast cancer survival using proportional hazards modeling and hierarchical clustering. The implications of aberrant ErbB expression have been explored by many previous investigations, and overall, HER-2 has been consistently associated with a shorter time to progression and decreased survival time, whereas correlative findings of the other ErbB receptors have varied widely.3-12 This is the first report of a quantitative protein detection method linking multiplexed ErbB expression to long-term patient outcome.
Most clinicians currently rely on clinicopathological parameters such as tumor size and nodal status, as well as ER, PR, and HER-2 tissue biomarkers to assess an individual's prognosis after surgery for primary breast cancer. We find that among the biomarkers measured, AQUA ER, AQUA PR, AQUA EGFR, AQUA HER-2, and AQUA HER-3 were significantly associated with long-term survival and independent of parameters of tumor size and nodal metastasis. We were unable to reproduce previous observations that HER-4 is a favorable prognostic biomarker.7, 8 In addition, the poor prognostic association of AQUA HER-2 and AQUA HER-3 were independent of other ErbB expression patterns. Clustering of the protein expression data revealed groups of breast cancer patients that coexpress sets of ErbB family members. Multiplexing of AQUA scores by hierarchical clustering classification was superior to conventional IHC or univariate AQUA classification of HER-2+ breast cancer for prognosis, and this effect is independent of current clinical staging variables.
This sort of clustering analysis has potential for use in classification of breast cancers in a manner similar to that done by cDNA array type studies.24 Although only 6 markers are used in this study, we included the 3 standard markers that are used in standard management of breast cancer (ER, PR, and HER-2), which allowed us to identify the triple negative subset of breast cancers. Examination of Figure 2 shows that the triple negative class self-assorts into Cluster IV. It is interesting to note that this study suggests that there are probably 2 biological classes within that group: the subset that is triple negative but expresses high levels of EGFR, and a second subset that is triple negative with high levels of HER-3. This observation confirms previous work reporting a high correlation between triple negative cases and EGFR overexpression.9, 25 Further studies are needed to assess the significance of the HER-3+ subdivision, but it could have implications for new ErbB-targeted therapies such as pertuzumab and cannertinib.26
Limitations of this study include its retrospective nature and the incompleteness of the treatment data. However, the collection and investigation of archival cohorts such as this allow valuable insights into the relationship of breast cancer outcome with the molecular features of primary tumors. These correlative studies suggest the investigation of multiplexed assessment of biomarkers as a method to predict response to therapy. In this study, assay conditions were carefully controlled using a specialized control cell and tissue microarray, which should ensure reproducibility in future studies now underway in cohorts treated with ErbB-targeted therapies. The results reported here show the power of quantitative protein-based multiplexed analysis. By collection of continuous scores proportional to protein expression of ErbB family members, we are able to define subsets of our cohort that show grouping that is analogous to cDNA-based classifications and is more specific and informative for prediction of outcome.
Conflict of Interest Disclosures
This work was supported by an Avon-NCI Progress for Patients grant and National Cancer Institute grants R33 CA 106,709 and R33 CA 110,511 to David L. Rimm and by a Medical Scientist Training Program grant to Jennifer M. Giltnane.
Robert L. Camp is a stockholder in, scientific founder of, and consultant to HistoRx, a private corporation to which Yale University has given exclusive rights to produce and distribute the software and technologies embedded in AQUA. Yale University retains patent rights for the AQUA technology.
David L. Rimm is a stockholder in, scientific founder of, and consultant to HistoRx.