in uence of biospecimen variables on proteomic biomarkers ......the correlation is moderate; between...

Biology of Human Tumors

Influence of Biospecimen Variables on ProteomicBiomarkers in Breast Cancer

Funda Meric-Bernstam1,2, Argun Akcakanat1, Huiqin Chen3, Aysegul Sahin4, Emily Tarco1, Selin Carkaci5,8,Beatriz E. Adrada5, Gopal Singh1, Kim-Anh Do3, Zerzhinski M. Garces1, Elizabeth Mittendorf2,Gildy Babiera2, Isabelle Bedrosian2, Rosa Hwang2, Savitri Krishnamurthy4, William F. Symmans4,Ana Maria Gonzalez-Angulo6,7, and Gordon B. Mills7

AbstractBackground:PI3K/Akt/mTOR signaling is being actively pursued as a therapeutic target for breast cancer.

We sought to determine if tumor heterogeneity andbiospecimen variables affect the evaluation of PI3K/Akt/

mTOR pathway markers.

Methods: Intraoperative image-guided core-needle biopsies (CNB), and central and peripheral surgical

tumor specimens were prospectively collected in 53 patients with invasive breast cancer. Specimens were

assessed with reverse-phase protein arrays (RPPA) and immunohistochemistry (IHC).

Results: There was a moderate or strong correlation between the expression of 149 (97%) of the 154

different RPPA markers in the center and periphery. Correlation was higher for smaller tumors, in patients

whodidnot undergoneoadjuvant therapy, andwith shorter cold ischemia time.Of 154markers, 132 (86%)

were not statistically different between the center and periphery, and 97 (63%) were not different between

the CNB and the surgical specimen (average of the central and peripheral specimen). pAkt S473 and PTEN

had a significant correlation between central and peripheral specimens, and between CNB and surgical

specimen. However, pAkt S473, pS6 S235/236, and pS6 240/244 levels were significantly higher in CNB

than the central specimens both by RPPA and by IHC.

Conclusions: Most individual proteomic biomarkers studied do not have significant intratumoral

heterogeneity. However, protein and phosphoprotein levels are affected by biospecimen type and other

preanalytic variables. PI3K pathway activation is greater in CNB compared with postexcision surgical

samples suggesting a potential loss of phosphorylation during surgicalmanipulation, or with cold ischemia

of surgical specimens. Clin Cancer Res; 20(14); 1–14. �2014 AACR.

IntroductionPI3K/Akt/mTOR signaling is being actively pursued as a

therapeutic target (1). However, assessment of PI3K/Akt/mTOR pathway activity remains a genuine obstacle topersonalizing therapy. Stability of phosphoproteins duringsample handling is amajor concern. In amulticenter trial incolorectal cancer, pAkt immunostaining was observed onlyin biopsies and not in surgical samples (2). Furthermore, inbreast cancer, significant loss in pAkt and pErk immunos-

taining was found in surgical specimens that underwentroutine processing compared with core-needle biopsies(CNB) taken from tumors immediately after excision (3).Thus, PI3K/mTOR activation status may vary based onspecimen acquisition approach and processing.

Fine-needle aspirate (FNA) and CNB are especially valu-able for decision-making for preoperative therapy of theprimary treatment na€�ve breast cancers, as well as in themetastatic setting. Previous work in our breast cancerresearch program demonstrated that FNAs enrich for tumorcells, retrieving 80% tumor cells compared with CNBs,which contained on average of 50% tumor cells (4). How-ever, how these tumor acquisition approaches affects pro-tein based biomarkers of PI3K/mTOR signaling have notbeen determined.

There is growing concern about tumor heterogeneity (5,6). Surgical excision allows assessment of the entire tumor,thus potentially overcoming concerns about tumor hetero-geneity. However, specimen quality is dependent on avariety of post-acquisition variables, including intraopera-tive ischemia, specimen handling, and time to preservation.Spruessel and colleagues have shown that initial changes ofgene and protein expression profiles were already observed

Authors' Affiliations: Departments of 1Investigational Cancer Therapeu-tics, 2Surgical Oncology, 3Biostatistics, 4Pathology, 5Radiology, 6BreastMedical Oncology, 7Systems Biology, The University of Texas MD Ander-sonCancerCenter, Houston, Texas; and 8Current InstitutionDepartment ofRadiology, Ohio State University, Columbus, Ohio

Note: Supplementary data for this article are available at Clinical CancerResearch Online (http://clincancerres.aacrjournals.org/).

Corresponding Author: Funda Meric-Bernstam, The University of TexasMD Anderson Cancer Center, 1400 Holcombe Boulevard, Unit 455, Hous-ton, TX 77030. Phone: 713-794-1226; Fax: 713-563-0566; E-mail:[email protected]

doi: 10.1158/1078-0432.CCR-13-1507

�2014 American Association for Cancer Research.

ClinicalCancer

Research

www.aacrjournals.org OF1

Research. on June 18, 2020. © 2014 American Association for Cancerclincancerres.aacrjournals.org Downloaded from

Published OnlineFirst June 3, 2014; DOI: 10.1158/1078-0432.CCR-13-1507

http://clincancerres.aacrjournals.org/

5 to 8 minutes after colon resection (4). The time topreservation may be especially prolonged in breast surgery,where excised specimens often undergo whole-specimenradiography followed by specimen margin inking and sec-tioning potentially with repeat radiography for marginassessment (7). However, in recent work, we demonstratedthat mRNA expression levels of single genes and multigenesignatures that are of diagnostic relevance in breast cancerwere mostly unaffected by sample preservation method orprolonged cold ischemic duration (8). In our preliminarywork with reverse-phase protein arrays (RPPA), we foundthat although the functional proteomic "fingerprint" wasreproducible in most breast cancer samples even after adelay of 24 hours before freezing, the expression of 21 of 82total and phosphoproteins was significantly affected bytime to tumor freezing, including markers of PI3K/mTORsignaling (pAkt T308, PTEN, pAMPK T172, TSC2, pTSC2T1462, S6K1, and pS6 S235/236; ref. 9), suggesting thatpostexcision expression of many proteins, especially phos-phoproteinsmay not accurately reflect the in vivo state of theproteins at the time of collection.

Immunohistochemistry (IHC) is the most commonlyused approach for assessing biomarkers at the protein levelbut is semiquantitative. RPPA has mainly been used as adiscovery tool, and has the advantage of being quantitative,allowing for assessment of hundreds of markers from asmall amount of tissue in a cost-effective fashion. However,unlike IHC, where the tumor versus stromal staining can bedifferentiated, in RPPA when total tumor lysate is used, thebiomarker signal will be more affected by low tumor cellu-larity and high heterogeneity. The ideal protein assessmentapproach may differ frommarker to marker, depending onbaseline tumor cellularity, marker expression pattern, andinfluence of sample handling.

We hypothesized that biospecimen variables and tumorheterogeneity may affect assessment of pathway activation.We therefore performed a prospective clinical study todetermine whether biospecimen variables including spec-imen type (in vivo collection of FNAor core biopsy vs. ex vivocollection of central and peripheral surgical tumor samples)affect the functional proteomic profile by RPPA and theassessment of PI3K/Akt/mTORpathway activation by RPPAand IHC. This work was performed to assess effect ofbiospecimen variables on RPPA results without microdis-section. We demonstrated that most biomarkers do nothave significant spatial heterogeneity in operable breastcancer. However, levels of several potentially actionablebiomarkers, including markers of PI3K pathway activation,are affectedbybiospecimen type andbiospecimenhandlingvariables.

Materials and MethodsHuman specimen collection

We conducted a prospective laboratory study afterapproval by the institutional review board (Fig. 1). Afterinformed consent, 53 patients with invasive breast cancer 2cm or greater on preoperative imaging, underwent intrao-perative ultrasound-guided percutaneous FNAs (4–8passes) and CNBs (3–4 core biopsies) after induction ofanesthesia but before the beginning of the operation. Thenumber of samples obtained per patient is listed in Sup-plementary Table S1. FNAs were frozen for RPPA. In addi-tion, one CNB with the greatest tumor content upon visualinspection was macrodissected and frozen for RPPA. Theother CNBs were formalin fixed and paraffin embedded(FFPE). The surgical specimen was evaluated by a dedicatedbreast pathologist in the frozen section suite. For breastconservation (lumpectomy) samples, whole specimenradiographs were performed for confirmation of targetlesion removal. For both breast conservation and mastec-tomy samples, margins were inked and sliced specimenswere radiographed as deemed necessary for margin analysisas previously described (7). A portion of central and periph-eral tumor was then allocated for research; one specimenfromeach sitewas snap frozen for RPPA, anotherwas placedin formalin. Central and peripheral research samples wereacquired at the same time, and the samples were approx-imately the same size, and were frozen directly. Smallspecimens of central and peripheral tumor samples wereremoved from main specimen in the frozen section suiteand then FFPE directly. Surgical tumor samples weremacro-dissected, microdissection was not performed. The timefrom intraoperative specimen removal to time of freezingof the research specimen was prospectively recorded anddesignated as "cold ischemia time."

The FFPE samples were assessed by hematoxylin andeosin (H&E) staining. Three patients had no tumor onH&Estaining of all 3 samples (CNB, central, and peripheralspecimen). These patients were excluded from further anal-ysis. Twelve patients had no tumor on H&E of 1 of the 3samples. The samples without tumor were excluded fromanalysis.

Translational RelevanceThere is great interest in using protein-based biomar-

kers, especially phosphoproteins for assessment of path-way activation, as potential predictors of response totargeted therapies. Our results demonstrate that mostindividual proteomic biomarkers did not have signifi-cant intratumoral heterogeneity. However, protein andphosphoprotein levels were affected by biospecimentype and other preanalytic variables. PI3K pathway acti-vation was greater in core-needle biopsies comparedwith surgical samples.Use of archival samples, especiallysurgical samples may introduce additional variability tobiomarker assessment. These findings have implicationsfor study design for biomarker-driven targeted therapytrials as well as window of opportunity studies. Thesestudies also highlight the need for biospecimen scienceto reduce postexcision degradation of phosphorylationas well as the need for identifying the most robustmarkers least prone to intratumoral heterogeneity andbiospecimen and assay-related variation.

Meric-Bernstam et al.

Clin Cancer Res; 20(14) July 15, 2014 Clinical Cancer ResearchOF2




Reverse-phase protein arraysRPPA was performed in the MD Anderson Cancer Center

Functional Proteomics Core Facility as described previously(10). Lysates were extracted from each tumor specimen and4 replicates from each lysate were spotted on the RPPA. Thecell lineswere treatedon separate days to obtain 3biologicalreplicates. Briefly, 200 mL of lysis buffer was added to CNBsamples. Following homogenization with ceramic beads,samples were centrifuged, and supernatant was quantitatedusing BCA Assay Kit (Thermo Scientific). The final proteinconcentration was adjusted to 2 mg/mL. FNA samples werelysed in 75 mL of lysis buffer that was alreadymixed with 25mL of loading buffer. Following sonication, samples werequantitated with BCA Assay Kit and the final concentrationwas adjusted to 3 mg/mL. Lysates were probed with anti-bodies validated for RPPA that are enriched for componentsof PI3K/Akt/mTOR pathway (Supplementary Table S2). Atotal of 154 unique proteins (including phosphoproteins)and 4 replicates were evaluated. Protein levels wereexpressed as the mean expression values in Log2. Sampleswith less than 20% tumor cellularity on correspondingH&Esections were not included in statistical analysis for RPPAdata.

ImmunohistochemistryIHC was performed on FFPE CNBs, and central and

peripheral tumor specimens with the following antibodies:pAkt S473 (Cell Signaling #4060, 1:50, 1 hour), pS6 S235/236 (Cell Signaling #4858, 1:50, overnight), pS6 S240/244(Cell Signaling #2215, 1:200, overnight), pPRAS40 T246

(Cell Signaling #2997, 1:200, overnight), and p4EBP1 T70(Cell Signaling #9455, 1:50, overnight) Ki67 and PTENimmunostaining was performed in the MD Anderson IHCCore laboratory with the following antibodies: Ki67 (Dako#M7240, 1:100, 20 minutes) and and PTEN (Dako#M3627, 1:100, 15minutes). Positive andnegative controlswere included for all assays. H-score was determined byestimation of the percentage of tumor cells positivelystained with weak, moderate, and strong staining intensity(by A. Sahin). The cytoplasmic score was used for statisticalanalysis. The following formula was used: H-score ¼(low%) � 1 þ (medium%) � 2 þ (high%) � 3.

Statistical analysisThe RPPA spot signal intensity data from MicroVigene

were processed using the R package SuperCurve (version1.4.3; ref. 1), available at "http://bioinformatics.mdander-son.org/OOMPA." RPPA superslide was normalized forFNA and surgical samples/CNB samples separately. Rawdata were treatedwithmedian centering across samples; thetreated data were undertaken with centering by the samplemedians; and the final normalized data are obtained byapplying MAD scaling to the data; the median centeredvalues were divided by median absolute deviation to scalethe data (11–13). For Spearman’s rank correlation test,P values are computed using algorithm Applied Statistics(AS) 89 (14, 15). The values of correlation were classified asfollows: rs of 0.8 to 1, the correlation is very strong; between0.6 and 0.79, correlation is strong; between 0.5 and 0.59,the correlation is moderate; between 0.3 and 0.49, the

Figure 1. Study schema. Patients underwent intraoperative breast biopsies with fine-needle aspirate (4–8 passes) and core-needle biopsy (3–4 passes). Thecore-needle biopsy with the greatest tumor cellularity was used for RPPA whereas the other core-needle biopsies were analyzed by IHC. After surgicalexcision, the specimen underwent routine intraoperative assessment including color-coded specimen margin inking, and sectioning and specimenmammogram as needed. Central and peripheral tumor specimen was collected for research to capture tumor heterogeneity. FNA, core-needle biopsy, andperipheral and central surgical resection samples underwent proteomic analysis with RPPA. Core-needle biopsy and both peripheral and centralsurgical resection samples underwent IHC for PI3K pathway markers.

Biospecimen Variables Affect Biomarkers

www.aacrjournals.org Clin Cancer Res; 20(14) July 15, 2014 OF3




correlation is moderate to low; between 0.16 and 0.29,correlation isweak to low; andbelow0.16, the correlation istoo low to be meaningful. Hierarchical Ward-linkage clus-tering was performed based on Spearman correlation coef-ficients. Paired t test was conducted to determine the dif-ference among RPPA data of various biospecimen types,andWilcoxon signed-rank test was applied to compare IHCdata. The Benjamini–Hochberg procedure was used tocontrol the false discovery rate (FDR). FDR of 0.05 wasconsidered significant.

For each of the 178 samples, we computed a PI3Kpathway score, which was the sumof the normalized values(median centering and MAD scaling) of the phosphopro-tein levels of Akt, mTOR, 4E-BP1, S6K, and S6 (i.e., PI3Kscore ¼ pS6 240/244 þ pS6 S235/236 þ pS6K T389 þpmTOR S2448 þ p4E-BP1 S65 þ p4E-BP1 T37/46 þpPRAS40 T246 þ pAkt S473 þ pAkt T308). We classifiedthe PI3K pathway scores of each sample type (FNA, CNB,central and peripheral surgical tumor) into 2 groups; sam-ples were considered activated if their PI3K scores were inthe top quartile within their sample type. Cohen k tests wereperformed to identify the agreement between PI3K pathwayactivity class and the biospecimen type. The k agreementwas characterized as values <0 as indicating no agreement, 0to 0.20 as slight, 0.21 to 0.40 as fair, 0.41 to 0.60 asmoderate, 0.61 to 0.80 as substantial, and 0.81 to 1 asalmost perfect agreement (16).

PTEN andpAktwere defined as loss and low, respectively,if their IHC values were 0. x2 tests were applied to check thedependency between PI3K score and PTEN loss/normal,PI3K score, and pAkt high/low. The analyses were per-formed by using R (version 2.11.1; R Development CoreTeam, 2011; ref. 17). Supplementary Table S3 describeshow each point of REMARK guidelines were addressed.

ResultsClinicopathologic characteristics

Clinical and pathologic characteristics of the patients arelisted in Supplementary Table S4. The median pathologictumor size was 2.8 cm. Most patients (77%) had hormone-receptor positive tumors. Eleven patients had receivedneoadjuvant chemotherapy; these patients had theirresearch biopsies intraoperatively (i.e., after completion ofthe neoadjuvant chemotherapy), followed by surgical exci-sion, similar to the surgery first cohort. Most patientsunderwent amastectomy. The specimen cold ischemia timewas an average of 37 minutes, and ranged from 8 to 157minutes.

Amedian of 6 FNA (range, 3–8) and 2 core-needle biopsy(range, 0–3) samples were obtained from each patient.Protein yield from a single FNA sample was a mean of1,635 mg (median yield 686 mg; range, 36–8,580 mg). Theprotein yield from a single core biopsy was a mean of 862.8mg (median yield 871.4 mg; range, 242–1,594 mg). Thetumor cellularity of each sample was assessed by H&E anddistribution of nonmalignant components in the centralspecimen including contribution of epithelium, lympho-cytes and endothelium, and fibroblasts was detailed (Sup-

plementary Table S5). There was no statistically significantdifference in tumor cellularity between the central andperipheral samples, but the tumor cellularity was lower inthe CNBs compared with the central specimens (in 47matched samples 54.8% vs. 61.2%; P ¼ 0.0689), andsignificantly lower in CNB compared with peripheral sam-ples (in 43matched samples 52.7% vs. 60.7%; P¼ 0.0476).Tumors whose corresponding FFPE samples had less than20% tumor cellularity or that only had ductal carcinoma insitu (DCIS) on FFPE were considered inevaluable for RPPA,except for studies comparisons on effect of tumor cellular-ity. Tumors whose IHC sections did not contain tumorwereconsidered inevaluable for IHC. Seven core biopsies, 4central specimens, and 4 peripheral specimens had lessthan 20% tumor cellularity on FFPE. Forty-four (88%) of50 CNBs, 46 (90%) of 51 of central tumor specimens, and43 (84%) of 51 peripheral tumor specimens were consid-ered evaluable for RPPA and IHC.

Effect of biospecimen type on proteomic markersTo determine the effect of biospecimen type on the

functional proteomic profile, we performed unsupervisedhierarchical clustering on the RPPA from FNAs, CNBs, andcentral and peripheral surgical samples. As demonstrated inthe heatmap (Fig. 2A), most of the FNA samples clusteredtogether, whereas the central and peripheral samples andCNB samples coclustered suggesting that the cellular con-tent of the FNA is markedly different from that of the CNBand surgical samples. Notably, CNB, central, and peripheralsamples from the same patient often clustered together,consistent with intertumoral differences being greater thanintratumoral differences. FNAs may not only differ becausethey are collected before any cold ischemia time, but alsobecause there are previously reported differences in relativetumor vs. stromal cell content (18), and technical differ-ences in RPPA sample processing and protein quantitationbetween FNA samples versusCNBand surgical samples. Thecorrelation between different biospecimens for each bio-marker is listed in Supplementary Table S6 anddepicted as aheatmap (Fig. 2B).

There was a moderate or strong correlation between thecentral and peripheral specimen results for all but 5 of 154proteins andphosphoproteins tested (Supplementary TableS6). There was a very strong correlation between the centraland peripheral samples for ER (r ¼ 0.896), PR (r ¼ 0.90),and strong correlation for HER2 (r¼ 0.786; Fig. 3). We alsocompared the central and peripheral protein expressionusing paired Student t test (Supplementary Table S7). In132 (86%)of 154markers tested, at a FDRof 0.05, therewasno statistical difference in the expression between centraland peripheral specimens.

To determine the impact of marker assessment on in vivobiopsies compared with ex vivo excision samples, we nextassessed whether RPPA results correlated between CNB andsurgical specimens. There was a very strong correlationbetween the CNB and surgical tumor specimen (average ofcenter and periphery) for ER (r ¼ 0.804), PR (r ¼ 0.871),and strong correlation with HER2 (r ¼ 0.66; Fig. 3). Thus,






our data suggest that for especially ER, PR, quantitation byRPPA does not show significant differences attributable tospatial intratumoral heterogeneity or biospecimen type.Therewas no statistically significant difference in expressionof 97 (67%) of the 154markers assessed in CNB, comparedwith the surgical samples by Student t test, at a FDR of 0.05(Supplementary Table S7).There was a modest to poor correlation between FNA

results and surgical specimens for several proteins (Supple-mentary Table S6). However, the FNA and surgical samplehad a significant correlation for 49 proteins, and the FNAand CNB had a correlation for 68 proteins. There wasstrong correlation between FNA and surgical samples forER (r¼0.606), PR (r¼0.695), and fair correlation forHER2(r ¼ 0.367). There was also a strong correlation betweenCNBand FNA results for ER (r¼ 0.650), PR (r¼ 0.642), andmoderate correlation for HER2 (r ¼ 0.57).

Effect of biospecimen variables on biomarkerexpressionWe next evaluated the effect of preanalytic biospecimen

variables on biomarker expression. At FDR < 0.05, in FNAsamples there was no difference in protein expressionbetween patients that were younger versus older thanmedi-an age of 51 years. In CNB, ER was higher in older patients(P < 0.0001), and TIGAR and c-Kit were higher in youngerpatients (P¼0.0008 and0.001, respectively). ER expression

also was significantly higher in central and peripheralsamples in patients that were older (P < 0.0001 for both).Therewas an increased pGSK3a/b S21/9 expression in grade3 tumors compared with grade 1 or grade 2 tumors (P ¼0.0343 and 0.0001, respectively). There was no proteindifferentially expressed between tumors of different stages,or between small versus large tumor size (median cut off 2.8cm) at FDR < 0.05 cut-off for any of the FNA, core, central,and peripheral samples.

We next determined whether preanalytic biospecimenvariables impacted the correlation of expression betweendifferent biospecimen types.We expected that larger tumorswould have greater tumor heterogeneity, and indeed bypaired t test, we found that correlation between central andperipheral specimenswas better in tumors that were smallerthan themean (<3.6 cm), comparedwith tumors larger thanthe mean (P < 0.001; Supplementary Table S8). Interest-ingly, larger tumors had a greater correlation between FNAversus surgical samples (P ¼ 0.003). Tumors with shorterthan themean cold ischemia time (36.7minutes) hadbettercorrelation between central versus peripheral specimenscompared with tumors with longer cold ischemia time(P < 0.001; Supplementary Table S8). There was also bettercorrelation between FNA and central specimens in tumorswith shorter cold ischemia time (P < 0.001). Shorter coldischemia time also was associated with a higher correlationbetween CNB and surgical specimens (P < 0.001).

Figure 2. RPPA on different biospecimens. A, unsupervised hierarchical clustering of RPPA results on different biospecimen types. Each column represents adifferent protein or phosphoprotein and each row, a different sample. Red, FNA samples; blue, core-needle biopsy; green, surgical samples. B, the correlationbetween different biospecimens for each biomarker depicted as a heatmap. Red, high positive correlation, green negative correlation.






We also assessed the effect of surgery type on biomarkercorrelation. There was a stronger correlation between CNBversus surgical samples (P < 0.001) as well as FNA vs.surgical samples (P ¼ 0.005) in patients who had breastconservation compared with those who underwent mas-tectomy. There was also a trend toward improved correla-tion between central and peripheral samples in breastconservation samples (P ¼ 0.074). To better understandwhether other preanalytic variables differed betweenbreast conservation andmastectomy samples,we compared

tumor size and cold ischemia time for these specimens.Therewas no statistically significant difference in tumor sizebetween patients who underwent breast conservation ver-sus mastectomy (median size 3.1 cm vs. 3.7 cm; P ¼ 0.49).However, surgery type was associated with differences incold ischemia time. Mean cold ischemia time for breastconservation patients was 28.5 minutes (range, 20–42min), compared with 39.1 minutes (range, 8–157 min) formastectomy (P ¼ 0.021). This may be attributable to thefact that breast conservation samples routinely undergo

Figure 3. Expression of PI3Kpathway markers by RPPAin different biospecimens.Correlation between ER, PR,HER2, pAkt S473, pAkt T308,PTEN, INPP4B, pS6 S235/236,pS6 S240/244. p4E-BP1 S65,p4E-BP1 T37/46, p4E-BP1 T70,pMAPK T202/204, and pMEK1/2S217/221 by RPPA betweendifferent biospecimens.






immediate margin analysis or maybe because mastectomyspecimens take longer to section.The correlation between FNA and surgical specimens was

significantly greater for nonphosphorylated proteins thanphosphorylatedproteins (P¼0.043). Therewas also a trendfor difference in correlation between CNB and surgicalspecimens (P ¼ 0.055). Thus, phosphoprotein levels aremore susceptible to biospecimen type/sample handlingthan total proteins.We next evaluated the effect of tumor cellularity on

biomarker expression (Supplementary Table S9). At FDR< 0.05, there are 54 proteins in CNB data showing signif-icant variability based on tumor cellularity. At FDR < 0.05,there were 22 proteins in central tumor data showingsignificant variability based on tumor cellularity. Therewere7 common proteins in CNB and central data that varied bytumor cellularity (cyclin B1, Beclin, IRS, YB1, p38 MAPK,BID, and eIF4G). The contribution of the tumor cell com-position was then correlated with the variation in theexpression of each marker between biospecimens. F test inlinear regression model was applied to check if the tumorcell composition affects the variance in the expression ofeach marker. The variance of most of the biomarkers couldnot be explained by the tumor cell composition.The biomarker correlation in the patients who did and

did not receive neoadjuvant chemotherapy was also ana-lyzed separately (Supplementary Table S10). Correlationsfor samples from patients who did not receive neoadjuvantchemotherapy were significantly higher than in patientswho received neoadjuvant chemotherapy when the com-paring the central versus peripheral specimen (P < 0.0001),CNB versus central specimen (P < 0.0001), CNB versussurgical specimen (P < 0.0001), FNA versus CNB (P <0.0001), and FNA versus peripheral sample (P ¼ 0.0209;Supplementary Table S11).

Effect of biospecimen typeonPI3K/Akt/mTORpathwayactivationAs thePI3K/Akt/mTORpathway is a potential therapeutic

target under active clinical investigation, we assessed PI3K/Akt/mTOR pathway activation by specimen type. First, weassessed expression of pAkt S473 and pAkt T308,markers ofactivated Akt (Fig. 3). pAkt S473 as measured by RPPA wascorrelated between all specimens and pAkt T308 was cor-related between the central versus peripheral and coreversus surgical specimen.PTEN and INPP4B are 2 tumor suppressors whose loss is

associated with activation of PI3K/Akt/mTOR signaling(19). The PTEN expression of central and peripheral sam-ples and CNB versus surgical samples were strongly corre-lated, both r¼ 0.716. The INPP4B expression of central andperipheral samples and CNB versus surgical samples werestrongly correlated, both r ¼ 0.774.Phosphorylation of mTOR target p4E-BP1 T37/46 corre-

lated strongly between central and peripheral samples(r ¼ 0.615), whereas p4E-BP1 S65 correlated moderately(r ¼ 0.531; Fig. 3). p4E-BP1 T37/46 and p4E-BP1 S65 hadmoderate-to-low correlation between the CNB and surgical

specimen (r ¼ 0.349 and 0.318, respectively). There wasalso a moderate-to-low correlation between central andperipheral samples for the mTOR downstream signalingmarkers pS6 S235/236 and pS6 240/244 (r ¼ 0.416 and0.439). The CNB and surgical specimens had no meaning-ful correlation between CNB and surgical specimen for pS6S235/236 (r ¼ 0.098) and weak-to-low correlation for pS6240/244 (r ¼ 0.25). The correlation between the CNB andFNA was somewhat stronger for pS6 S235/236 (r ¼ 0.364)and pS6 240/244 (r ¼ 0.415).

As the concordance of PI3K pathway activation in theCNB and surgical specimens was moderate to low, we nextdetermined whether there was a significant difference inPI3K pathway activation in CNB versus surgical specimenby paired Student t test. When the CNB was compared withthe central specimen, expressionwashigher inCNB for pAktS473 (P¼ 0.020), p4E-BP1 S65 (P¼ 0.025), p4E-BP1 S37/46 (P < 0.001), pS6 S235/236 (P ¼ 0.011), and pS6 S240/244 (P < 0.001), as well as PTEN levels (P <0.001) andpMAPK T202/204 (P ¼ 0.004; Fig. 4). When we comparedwith the overall surgical specimen (average of the center andperiphery), p4E-BP1 T37/46 and pS6 240/244 levels in theCNB were significantly higher (P ¼ 0.005 and P < 0.001,respectively).

PI3K/Akt/mTOR signaling by IHCWe determined PI3K/Akt/mTOR signaling using IHC.

Examples of positive and negative staining for each anti-body are shown in Fig. 5A.

There was a very strong correlation in p4E-BP1 T70expression in the central and peripheral tumor (r ¼0.81). There was a moderate correlation in central andperipheral expression of pAkt S473 (i ¼ 0.588), pPRAS40T246 (r¼ 0.82), pS6 S240/244 (r ¼ 0.553), and moderate-to-low correlation for pS6 S235/236 (r ¼ 0.475; Fig. 5B).There was a very strong correlation in PTEN expression inthe central and peripheral tumor (r ¼ 0.85; Fig. 5B). WhenPTEN IHC was categorically classified as having PTEN lossversus not, the central and peripheral samples had almostperfect agreement (k ¼ 0.916), CNB and central specimenhad substantial agreement (k ¼ 0.682), and CNB andperipheral specimen had moderate agreement (k ¼ 0.579).

The CNB and central specimen immunostaining alsocorrelated for pAkt S473 (r ¼ 0.87), PTEN (r ¼ 0.72),p4E-BP1 T70 (r ¼ 0.81), and pS6 S235/236 (r ¼ 0.38), butnot for pPRAS40 T246 or pS6 S240/244. The CNB andperipheral specimen immunostaining were correlatedfor pAkt S473 (r ¼ 0.71), PTEN (r ¼ 0.68), p4E-BP1 S65(r ¼ 0.63), pS6 S240/244 (r ¼ 0.47), and pS6 S235/236(r ¼ 0.49).

Wenext compared PI3Kpathway activation status by IHCto that of RPPA. pAkt S473 quantitation by RPPA correlatedwith immunostaining by IHC in central specimen(r ¼ 0.71), peripheral specimen (r ¼ 0.65), and CNB(r ¼ 0.52; Fig. 5C). RPPA quantitation of PTEN correlatedwith IHC results in the central specimen (r¼ 0.41), but notin the peripheral tumor or CNB specimens. pS6 S235/236expression by RPPA and IHC also were significantly






correlated in the central specimen and in the CNB, but nottumor periphery (r ¼ 0.24). When RPPA quantitation ofFNAs was compared with IHC of central specimens, therewas moderate correlation for pAkt S473 (i¼ 0.30), PTEN (i¼ 0.37), and pS6 S235/236 (r ¼ 0.35).

There was a very strong correlation between percentagecells positive for Ki67 between the central and peripheralspecimen (r ¼ 0.865) and between the CNB and centralspecimen (r ¼ 0.891). There was no significant correlationbetween the expression of proliferation marker proliferat-ing cell nuclear antigen (PCNA) on RPPA and percentagecells positive for Ki67 on IHC.

We then determined whether there were differences inimmunostaining by biospecimen type. The pAkt S473

expression in the CNB was significantly higher than thatin the central specimen (P ¼ 0.013; Fig. 6A) but not theperipheral specimen. The pS6 S240/244 and pS6 S235/236H-scores on theCNBwere significantly higher than both thecentral specimen and the peripheral specimen (P<0.001 forboth; Fig. 6A). Examples of discordant IHC results withhigher immunostaining in core biopsy than surgical speci-mens are shown in Fig. 6B.

Concordance of PI3K pathway classification by RPPAWe assessed pAkt S473, pAkt T308, pPRAS40 T246, p-

mTOR S2448, p4E-BP1 S65, p4E-BP T37/46, pS6K T389,pS6 240/244, and pS6 S235/236 levels as indicators of highPI3K pathway activity, as they are known downstreamtargets (10, 20), and as their phosphorylation was foundto be strongly inhibited by PI3K/mTOR inhibitor BEZ235(100 nM) and Akt inhibitor MK-2206 (10 mM) in a panelof cancer cell lines (Fig. 7A and B).

When all samples were analyzed together, as expected,tumors with expression of pAkt S473 on IHC also had asignificantly higher PI3K score than those that did not (P¼0.001; Fig. 7C). Furthermore, samples with PTEN loss byIHC (H-score¼ 0) had a significantly higher PI3K pathwayactivation score (P ¼ 0.018; Fig. 7C). The statistical differ-ence in PI3K activity score was not preserved when thethreshold for PTEN loss was elevated to H-score of 50 or100, suggesting that complete PTEN loss establishes thegreatest differentiation between tumors that have PI3Kactivation versus not. When PTEN IHC and PI3K activityon RPPA was analyzed separately for each sample type(core, central, and peripheral), PI3K activity score wassignificantly higher in CNB that had PTEN loss versus thosewithout PTEN loss by IHC (P¼ 0.0104), whereas there wasno there was no difference in PI3K activity score central orperipheral samples that had PTEN loss versus those withoutPTEN loss by IHC (Fig. 7D).

We then classified tumors as "PI3K signaling activated" or"not activated" based on PI3K activity score calculatedseparately for each specimen type. F test in linear regressionmodel demonstrated that the variance of PI3K pathwayscore could not be explained by the tumor cell composition.

The samples in the top quartile of PI3K RPPA core withineach specimen type were classified as "PI3K signaling acti-vated." In the 37 patients where RPPA results were availableon both CNB and central specimens, the overall agreementbetween the core biopsy and central specimenwas 78.4% (k¼ 0.454, P ¼ 0.005). In the 39 patients for whom bothcentral and peripheral RPPA results were available, theoverall agreement was 74.4% (k ¼ 0.33; P ¼ 0.04). Forsamples that had all 4 biospecimen types were available,activity of PI3K pathway for each RPPA biospecimen type isshown in Fig. 7E.

DiscussionTo deliver on the promise of personalized therapy, it is

critical to develop robust markers that can assess activationof druggable targets such activation of PI3K/Akt/mTOR

Figure 4. Comparison of expression in core-needle biopsy and centralspecimens. ER, PR, pAkt S473, pAkt T308, pAkt T308, PTEN, pS6S235/236, pS6 S240/244, p4E-BP1 S65, p4E-BP1 S37/46, and pMAPKT202/204 expression in the core-needle biopsy and central specimenby RPPA.






signaling in patient samples. Furthermore, in early clinicaltrials with targeted therapies, it is critical to incorporatepharmacodynamic studies to determine whether a newdrug inhibits its target and by how much. Such studies relyon the ability to accurately assess proteomic markers intumor samples. Marker assessment for patient selection isoften performed on archival surgical tissue in FFPE.However, there are significant concerns about tumor het-erogeneity and marker stability, especially for phospho-proteins. Therefore, we sought to determine how tumorheterogeneity and biospecimen variables affect assessment

of proteomic biomarkers and, in particular, PI3K/Akt/mTOR pathway activation. Proteomic levels were differen-tially affected by biospecimen type and other preanalyticvariables. For example, PI3K pathway activation was greaterin CNB compared with surgical samples suggesting loss ofphosphorylation during tumor handling during surgeryand cold ischemia after removal from the patient.

In genomic studies, intratumoral heterogeneity wasobserved for mutations in oncogenes and multipletumor-suppressor genes, even with gene-expression signa-tures of both good and poor prognosis being detected in

Figure 5. IHC for PI3K markers.A, IHC for PTEN, pAkt S473,pPRAS40 T246, p4E-BP1 S65,pS6 S240/244, and Ki67.Examples of positive (top) andnegative (lower) immunostainingfor each marker are included. B,correlation of immunostaining forPTEN, pAkt S473, pS6 S240/244,andpS6S235/236between centraland peripheral specimens (left),and central specimens andcore-needle biopsy (right).Immunostaining was quantitatedby H-score. C, correlation ofcentral RPPA results with core-needle biopsy specimenimmunostaining for PTEN, pAktS473, pS6 S240/244, and pS6S235/236.






different regions of the same tumor (6). Thus, we wereconcerned that there may be significant heterogeneity inproteomic markers, and we hypothesized that there may bedifferences between the tumor center andperiphery becauseof tumor hypoxia, etc. Indeed, the PI3K pathway classifica-tion by RPPA activation score did differ for some tumors.However, there was no statistical difference in the expres-sion between central and peripheral specimens for most(86%) of the individual markers tested and there was amoderate-to-strong correlation between the central andperipheral specimen results for all but 5 markers tested.The correlation between central and peripheral sampleswasstronger in smaller tumors suggesting there may be greatertumor heterogeneity in larger tumors.

There is a great interest in PI3K/Akt/mTOR signaling. Ourstudies suggest that pAkt on RPPA and PTEN on IHC arerelatively robust markers. As Akt phosphorylation is corre-lated with kinase activity, pAkt is one the most compellingmarkers to assess forpathway activation. pAkt expressiononRPPA in different biopsy and surgical biospecimens showedamoderate-to-strong correlation. pAkt staining on IHC hada more limited range; however, pAkt expression in RPPAand IHC showed a moderate correlation, further strength-ening the reliability of pAkt as a marker of pathway activity.The tumor suppressor PTEN is a critical negative regulator ofPI3K signaling that is aberrantly expressed in many tumortypes, and thus is already being used as a marker for patientselection in ongoing clinical trials for PI3K/Akt/mTORinhibitors. Our results show that PTEN IHC has a verystrong correlation between central and peripheral speci-mens and between CNB and surgical specimens. The PTENexpression of central and peripheral samples and CNBversus surgical samples were strongly correlated; however,the correlation of PTEN expression between RPPA and IHCwas weak. This is likely because of the effect of stroma onRPPA results. Even if PTEN is lost in tumor cells, PTENexpression in stroma is retained, and in fact stroma serves asa positive control in IHC. On IHC, only tumor PTENstaining is scored, but on RPPA, in the absence of micro-

dissection, quantitation of PTEN in the tumor lysate repre-sents PTEN expression in stroma as well as tumor cells.Thus, in less cellular tumors PTEN loss may not be asapparent on RPPA, and this may contribute to weak corre-lation between IHCandRPPA. PTEN IHC is attractive as notonly does it seem to have low interspecimen variability, butalso PTEN loss and inparticular complete loss (H-score¼0)on IHC was associated with PI3K pathway activation byRPPA. Larger studies are needed to further test the impact on"low" PTEN on pathway activation in order to set theappropriate threshold for PTEN loss that is biologically andclinically meaningful.

There has been increasing concerns that specimen type(such as CNB vs. surgical sample) may affect biomarkers.When standard breast cancer markers in CNB were com-pared with excisions, discordance between CNB and exci-sion for HER2 and ER expression was found to be rare(1.8% and 1.2%, respectively), however discordance in PRexpression was common (15%; ref. 21). A recent meta-analysis that included 21 articles involving nearly 2,500patients compared concordance of ER and PR in CNBs andexcision and found an overall agreement between CNBand surgical excisions of 92.8% for ER (k ¼ 0.78) and85.2% for PR (k ¼ 0.66; ref. 22). In contrast, Baker andcolleagues reported that, in colorectal cancer, pAkt S473staining was observed only in biopsies and not in surgicalsamples (albeit not assessed with RPPA; ref. 2), thusauthors suggested that caution should be used when usingphosphoprotein levels in surgical specimens to measureintrinsic signaling activity or drug effects because of thepotential for rapid dephosphorylation. Recently, to eval-uate the impact of ischemia, Mertins and colleagues col-lected ovarian tumor and breast cancer xenograft andperformed quantitative proteomics and phosphoproteo-mics with RPPA and liquid chromatography/mass spec-troscopy after defined ischemic intervals (23). Althoughthe global expressed proteome and most of the >25,000quantified phosphosites were unchanged after 60 minutes,rapid phosphorylation changes were observed in up to

pS6 S235/236pS6 S240/244pAkt S473

P = 0.0133 P < 0.001P < 0.001

A

S473pAkt

pS6 S240/244

pS6 S235/236

Core Central PeripheralB

Core Central Core Central Core Central

050

100

150

200

250

300

050

100

150

200

250

300

050

100

150

200

250

300

pS

6 S

235/

236

pA

kt S

473

pS

6 S

240/

244

Figure 6. Comparison of IHC incore biopsy and central surgicalspecimens. A, immunostaining forpAkt S473, pS6S240/244, andpS6S235/236 in matched core biopsyand central surgical specimens.IHC immunostaining wasquantitated with H-score. B,examples of discordant sampleswith strong immunostaining forpAkt S473, pS6S240/244, andpS6S235/236 in the core biopsy, butnot in the matched central orperipheral specimens.






24% of the phosphoproteome. Concern about stability ofcell signaling markers was also raised in a recent study onbreast cancer, in which loss of pAkt and pErk was found byIHC on CNB obtained ex vivo but not in surgical samplesafter routine processing (3). These studies suggest that lossof phospho-staining can occur during routine fixation ofresection specimens. Our study design was different thanthe latter breast cancer study as our CNB were obtained noton surgical samples ex vivo, but in the operating room invivo, while samples of surgical specimens were frozen or

formalin fixed immediately upon specimen processing forclinical care, thus mimicking earliest potential time pointfor specimen collection for tumor banking or biomarkerassessment. Our study findings are consistent with theprevious reports in that pAkt and pS6 levels were higherin CNB, and there were several tumors in which the PI3Kscore was elevated in the CNB, but not in the surgicalsample. Taken together, our results suggest that PI3Kpathway activation is greater in vivo, with subsequentdephosphorylation/pathway inactivation ex vivo.

Figure 7. Activity of PI3K pathwayby RPPA sample type. Regulationof PI3K downstream signaling byPI3K/mTOR inhibitor BEZ235 (A)and Akt inhibitor MK-2206 (B).C, PI3K activation score by IHCresults. PI3K score in patients withloss of PTENon IHC (PTENHscoreof 0) vs. PTEN H score > 0 (left).PI3K score in patients with lowpAkt S473 on IHC (pAkt S473 Hscore of 0) vs. pAkt S473H-score >0 (right). D, PI3K activation bybiospecimen type. Activatedsamples byRPPAPI3K score are inred, not activated samples ingreen.Only patientswith all sampletypes available are presented inthis figure. E, PI3K pathway activityby RPPA for samples that had all 4biospecimen types was available.Red activated, green not activated.






PI3K/Akt/mTOR pathway activity is being pursued as apotential predictor of response, marker of target inhibition,and pharmacodynamic marker of response. Recently, in arandomized, double-blind, placebo-controlled, phase IIItrial, Jerusalem and colleagues reported that patients withHER2þmetastatic breast cancerwith high pS6 levels derivedmore benefit from the addition of everolimus to trastuzu-mab and vinorelbine (24). However, our data and thosefrom others suggest results of PI3K/Akt/mTOR testing suchas pS6 levels may vary based on specimen acquisitionapproach and processing. This has several clinical implica-tions. First, PI3K activation is likely better assessed on CNBthan surgical specimens. Thus, if pathway activation is to bepursued as a potential prognostic or predictive biomarker,CNBs may be preferred over surgical specimens for bio-marker assessment. Expedited biospecimen processing/preservation approaches for surgical specimens may alsoimprove accuracy. Second, this finding has study designimplications for window of opportunity and neoadjuvanttherapy studies planned with PI3K inhibitors. In thesestudies, breast tumors are treated with PI3K/Akt/mTOR-targeted therapy for weeks to months after a pretreatmentbiopsy to assess effects on pathway activation comparedwith baseline levels. As PI3K activation may be higher inCNB compared with surgical samples even without treat-ment, pharmacodynamic effects of targeted therapies couldbebest assessed either byobtaining aposttreatmentCNB forcomparison to pretreatment CNB or by having a placebocontrolled arm to allow for comparison of surgical speci-mens in both arms.

It is encouraging that there was a moderate/strongcorrelation between the CNB and surgical specimen formost proteins and phosphoproteins. The expression ofseveral putative prognostic markers and druggable targetscorrelated significantly between the CNB and the surgicalspecimen, but some did not. Thus, it is important todetermine effect of biospecimen type for each markerused to determine its limitations. It may then be possibleto identify a set of biomarkers that can be robustlymeasured on both CNB and surgical specimens obviatingsome of the concerns with sample handling. The corre-lation of CNB with surgical specimen was greater thanthat seen between FNA and surgical specimen. This is notsurprising given that FNA is likely to be enriched forepithelial cells and have fewer stroma cells, and perhapsa larger contribution from blood. However, it is notablethat several markers were significantly correlated betweenFNA and surgical sample as well, including standard ofcare markers such as ER, PR, and HER2. Thus, truly robustmarkers are likely to be reliable regardless of biospecimentype. Indeed, ER, PR, and HER2 all tend to have a largedynamic range and to also be dichotomous markerssuggesting that these variables may contribute to therobustness of the marker. This is further highlighted bythe fact that several studies have successfully used FNAs todemonstrate the prognostic role of known and novelproteomic markers and to assess pharmacodynamicsmarkers of response to targeted therapies (10, 25).

Our study has several limitations. In our study, all FNAsand core biopsies were done intraoperatively, and wereimmediately followed by surgical resection. However, FNAand core-needle biopsies are invasive procedures that mayresult in wounding/bleeding/tissue damage. The localwound response itself may influence cell signaling, thusaffecting proteomic profile.

Furthermore, as we collected surgical samples in thefrozen section suite at the first possible time point, andsmall portions of tumor were formalin fixed, we are notcapturing thediscordance thatwouldhave emergedbecauseof variation in fixation of larger specimens. Concordancebetween biospecimens has been variable, suggesting thatpreanalytic variablesmay affectmarkers differentially; how-ever, in the absence of clinical outcome data, the impact ofthis variability cannot be assessed and thresholds for accept-able interspecimen heterogeneity has not been establishedfor most markers. Before clinical transition of any biomark-er, interassay reproducibility needs to be well establishedand clinically relevant cut-offs for each marker needs to bedetermined. In our analysis, correlation of biomarkers indifferent biospecimens varied from being strongly correlat-ed to not being correlated. Furthermore, discrepancies wereseen between different platforms and specimens in bothdirections. Biomarkers that have less intraspecimen vari-ability are likely to be more robust, more reproduciblebiomarkers with greater clinical utility. RPPA is a compel-ling research tool as it is tissue sparing, and allows forassessment of multiple markers from a small amount ofprotein lysate. One limitation of RPPA of macrodissectedtumor lysate is that it represents the signal in the tumor cellsas well as the surrounding stroma. An area of controversyhas been whether laser-aided microdissection is required(26). We performed RPPA after macrodissection, withoutlaser capture microdissection, an approach also used in TheGenome Cancer Atlas for breast cancer as well as severalother cancer types (26–28). As we did not do a side by sidecomparisonofmicrodissected andmacrodissected samples,in our study we did not test the utility of microdissection.Although IHC allows for assessment of fewer markers andmay be less quantitative, it is a more readily availableapproach, and provides information about localization ofexpression, and stromal or other normal contamination isless of a concern. Other biospecimen differences such assample type may affect results of both assay types. Optimalbiospecimen type and assay approach should be consideredfor each biomarker and application.

In summary, overall there is limited spatial intratumoralheterogeneity for individual proteomicmarkers in operablebreast cancer. However, some actionable markers are dif-ferentially expressed by biospecimen type and other pre-analytic variables. Although expression of most markers onCNB and surgical samples correlate, some do differ. Nota-bly, PI3K pathway activation is greater in CNB comparedwith surgical samples, thus CNB may be preferable tosurgical samples to assess pathway activity. Careful studyof preanalytic and analytic variables that influence results isneeded for clinical implementation of each biomarker.






Disclosure of Potential Conflicts of InterestG.B.Mills reports receiving commercial research grants fromAstraZeneca,

Celgene, CeMines, Exelesis/Sanofi Aventis, GlaxoSmithKline, Roche, SDI,andWyeth/Pfizer/Puma; has ownership interests in Catena Pharmaceuticals,PTV Ventures, Spindle Top Ventures; and is a consultant/advisory boardmember for Arcxis Biotechnologies, Asuragen, Aushon, Catena Pharmaceu-ticals, Daiichi Pharmaceuticals, Foundation Medicine, Han AlBio Korea,Komen Foundation, Novartis, Targeted Molecular Diagnostics, and TauTherapeutics. No potential conflicts of interest were disclosed by the otherauthors.

Authors' ContributionsConception and design: F. Meric-Bernstam, K.-A. Do, G. Babiera,S Krishnamurthy, A.M. Gonzalez-Angulo, G.B. MillsDevelopment of methodology: F. Meric-Bernstam, A. Sahin, K.-A. DoAcquisitionofdata (provided animals, acquired andmanagedpatients,provided facilities, etc.): F. Meric-Bernstam, A. Akcakanat, S. Carkaci, B.E.Adrada, G. Singh, Z.M. Garces, E. Mittendorf, I. Bedrosian, W.F. Symmans,G.B. MillsAnalysis and interpretation of data (e.g., statistical analysis, biosta-tistics, computational analysis): F. Meric-Bernstam, H. Chen, K.-A. Do,G. Babiera, R. Hwang, A.M. Gonzalez-AnguloWriting, review, and or revision of the manuscript: F. Meric-Bernstam,K.-A. Do, E. Mittendorf, G. Babiera, I. Bedrosian, R. Hwang, S. Krishna-murthy, W.F. Symmans, A.M. Gonzalez-Angulo, G.B. Mills

Administrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): F. Meric-Bernstam, A. Sahin,E. Tarco, G. Singh, Z.M. Garces, G.B. MillsStudy supervision: F. Meric-Bernstam, K.-A. DoOther (played role in tissue recruitment, performed the core-needlebiopsies and fine-needle aspiration biopsies): S. Carkaci, B.E. Adrada

Grant SupportThis work was supported by a Society of Surgical Oncology Clinical

Investigator Award in Breast Cancer Research sponsored by Susan G.Komen for the Cure (F. Meric-Bernstam), Susan G. Komen for the CureSAC10006 (F. Meric-Bernstam, K.-A. Do), Stand Up to Cancer DreamTeam Translational Research Grant, a Program of the EntertainmentIndustry Foundation administered by the American Association forCancer Research (SUC2-AACR-DT0209) (A.M. Gonzalez-Angulo, G.B.Mills, A. Akcakanat, F. Meric-Bernstam), the Kleberg Foundation (F.Meric-Bernstam, A.M. Gonzalez-Angulo, G.B. Mills), and NCRR Grant3UL1RR024148 (F. Meric-Bernstam, K.-A. Do) and UL1TR000371 (A.Akcakanat, F. Meric-Bernstam, K.-A. Do), and the National CancerInstitute through R21CA159270 (F. Meric-Bernstam, A.M. Gonzalez-Angulo, E. Tarco) and The University of Texas MD Anderson CancerCenter Support Grant (P30 CA016672).

Received May 31, 2013; revised April 25, 2014; accepted May 13, 2014;published OnlineFirst June 3, 2014.

References1. McAuliffe PF, Meric-Bernstam F, Mills GB, Gonzalez-Angulo AM.

Deciphering the role of PI3K/Akt/mTOR pathway in breast cancerbiology and pathogenesis. Clin Breast Cancer 2010;10 Suppl 3:S59–65.

2. Baker AF, Dragovich T, Ihle NT, Williams R, Fenoglio-Preiser C, PowisG. Stability of phosphoprotein as a biological marker of tumor signal-ing. Clin Cancer Res 2005;11:4338–40.

3. Pinhel IF, Macneill FA, Hills MJ, Salter J, Detre S, A'Hern R, et al.Extreme loss of immunoreactive p-Akt and p-Erk1/2 during rou-tine fixation of primary breast cancer. Breast Cancer Res 2010;12:R76.

4. Spruessel A, Steimann G, Jung M, Lee SA, Carr T, Fentz AK, et al.Tissue ischemia time affects gene and protein expression patternswithin minutes following surgical tumor excision. BioTechniques2004;36:1030–7.

5. Yachida S, Jones S, Bozic I, Antal T, Leary R, Fu B, et al. Distantmetastasis occurs late during the genetic evolution of pancreaticcancer. Nature 2010;467:1114–7.

6. Gerlinger M, Rowan AJ, Horswell S, Larkin J, Endesfelder D,Gronroos E, et al. Intratumor heterogeneity and branched evolutionrevealed by multiregion sequencing. N Engl J Med 2012;366:883–92.

7. Cabioglu N, Hunt KK, Sahin AA, Kuerer HM, Babiera GV, SingletarySE, et al. Role for intraoperative margin assessment in patientsundergoing breast-conserving surgery. Ann Surg Oncol 2007;14:1458–71.

8. Hatzis C, Sun H, Yao H, Hubbard RE, Meric-Bernstam F, Babiera GV,et al. Effects of tissue handling on RNA integrity and microarraymeasurements from resected breast cancers. J Natl Cancer Inst2011;103:1871–83.

9. Hennessy BT, Lu Y, Gonzalez-Angulo AM, Carey MS, Myhre S, Ju Z,et al. A technical assessment of the utility of reverse phase proteinarrays for the study of the functional proteome in non-microdissectedhuman breast cancers. Clin Proteom 2010;6:129–51.

10. Meric-Bernstam F, Akcakanat A, Chen H, Do KA, Sangai T, Adkins F,et al. PIK3CA/PTEN mutations and Akt activation as markers ofsensitivity to allosteric mTOR inhibitors. Clin Cancer Res 2012;18:1777–89.

11. Dudoit S, Yang YH. Bioconductor R packages for exploratory analysisand normalization of cDNA microarray data. In:Parmigiani G, GarrettES, Irizarry RA, Zeger SL, editors. The analysis of gene expressiondata: methods and software. New York: Springer; 2002.

12. Yang YH, Dudoit S, Luu P, Speed TP. Normalization for cDNA micro-array data. In:Bittner ML, Chen Y, Dorsel AN, Dougherty ER, editors.Microarrays: optical technologies and informatics: SPIE—InternationalSociety for Optical Engine; 2001.

13. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, Ngai J, et al. Normalizationfor cDNA microarray data: a robust composite method addressingsingle and multiple slide systematic variation. Nucleic Acids Res2002;30:e15.

14. Best DJ, Roberts DE. Algorithm AS 89. The upper tail probabilities ofSpearman's rho. Appl Stat 1975;24:377–79.

15. HollanderM,Wolfe DA. Nonparametric statistical methods. New York:John Wiley & Sons; 1973.

16. Landis JR, Koch GG. The measurement of observer agreement forcategorical data. Biometrics 1977;33:159–74.

17. Hu J, He X, Baggerly KA, Coombes KR, Hennessy BT, Mills GB. Non-parametric quantification of protein lysate arrays. Bioinformatics(Oxford, England) 2007;23:1986–94.

18. Symmans WF, Ayers M, Clark EA, Stec J, Hess KR, Sneige N, et al.Total RNA yield and microarray gene expression profiles from fine-needle aspiration biopsy and core-needle biopsy samples of breastcarcinoma. Cancer 2003;97:2960–71.

19. Gewinner C,Wang ZC, Richardson A, Teruya-Feldstein J, Etemadmo-ghadam D, Bowtell D, et al. Evidence that inositol polyphosphate 4-phosphatase type II is a tumor suppressor that inhibits PI3K signaling.Cancer Cell 2009;16:115–25.

20. Meric-Bernstam F, Gonzalez-Angulo AM. Targeting the mTORsignaling network for cancer therapy. J Clin Oncol 2009;27:2278–87.

21. Arnedos M, Nerurkar A, Osin P, A'Hern R, Smith IE, Dowsett M.Discordance between core needle biopsy (CNB) and excisionalbiopsy (EB) for estrogen receptor (ER), progesterone receptor (PgR)and HER2 status in early breast cancer (EBC). Ann Oncol 2009;20:1948–52.

22. Li S, Yang X, Zhang Y, Fan L, Zhang F, Chen L, et al. Assessmentaccuracy of core needle biopsy for hormone receptors in breastcancer: a meta-analysis. Breast Cancer Res Treat 2012;135:325–34.

23. Mertins P, Yang F, Liu T, Mani DR, Petyuk VA, Gillette MA, et al.Ischemia in tumors induces early and sustained phosphorylationchanges in stress kinase pathways but does not affect globalprotein levels. Mol Cell Proteomics 2014 May 7. [Epub ahead ofprint]






24. Jerusalem G, Andre F, Chen D. Evaluation of everolimus (EVE) inHER2þ advanced breast cancer (BC) with activated PI3K/mTOR path-way: exploratory biomarker observations from the BOLERO-3 trial.European Cancer Congress. Amsterdam, the Netherlands; 2013. p.Abstract LBA16.

25. Gonzalez-Angulo AM, Hennessy BT, Meric-Bernstam F, Sahin A, LiuW, Ju Z, et al. Functional proteomics can define prognosis and predictpathologic complete response in patients with breast cancer. ClinProteomics 2011;8:11.

26. Mueller C, Decarvalho AC, Mikkelsen T, Lehman NL, Calvert V, EspinaV, et al. Glioblastoma cell enrichment is critical for analysis of phos-phorylated drug targets and proteomic-genomic correlations. CancerRes 2014;74:818–28.

27. Cancer GenomeAtlas Network. Comprehensive molecular portraits ofhuman breast tumours. Nature 2012;490:61–70.

28. Akbani R, Ng PKS,Werner H, Shahmoradgoli M, Zhang F, Ju Z, et al. Apan-cancer proteomic perspective on The Cancer Genome Atlas. NatCommun 2014; 29:3887.






Published OnlineFirst June 3, 2014.Clin Cancer Res Funda Meric-Bernstam, Argun Akcakanat, Huiqin Chen, et al. Biomarkers in Breast CancerInfluence of Biospecimen Variables on Proteomic

Updated version

10.1158/1078-0432.CCR-13-1507doi:

Access the most recent version of this article at:

Material

Supplementary

http://clincancerres.aacrjournals.org/content/suppl/2014/07/18/1078-0432.CCR-13-1507.DC1Access the most recent supplemental material at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected] at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

Permissions

Rightslink site. (CCC)Click on "Request Permissions" which will take you to the Copyright Clearance Center's

.http://clincancerres.aacrjournals.org/content/early/2014/06/30/1078-0432.CCR-13-1507To request permission to re-use all or part of this article, use this link



http://clincancerres.aacrjournals.org/lookup/doi/10.1158/1078-0432.CCR-13-1507

http://clincancerres.aacrjournals.org/content/suppl/2014/07/18/1078-0432.CCR-13-1507.DC1

http://clincancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://clincancerres.aacrjournals.org/content/early/2014/06/30/1078-0432.CCR-13-1507


in uence of biospecimen variables on proteomic biomarkers ......the correlation is moderate; between...

Documents