Cancer Diagnosis Program | Biorepositories and Biospecimen Research Branch

Search BRD Papers

Influence of biospecimen variables on proteomic biomarkers in breast cancer.

Author(s): Meric-Bernstam F, Akcakanat A, Chen H, Sahin A, Tarco E, Carkaci S, Adrada BE, Singh G, Do KA, Garces ZM, Mittendorf E, Babiera G, Bedrosian I, Hwang R, Krishnamurthy S, Symmans WF, Gonzalez-Angulo AM, Mills GB

Publication: Clin Cancer Res, 2014, Vol. 20, Page 3870-83

PubMed ID: 24895461 PubMed Review Paper? No

Purpose of Paper

Conclusion of Paper

Studies

1. Study Purpose

   The purpose of this study was to compare protein profiles of case-matched specimens obtained by FNA, CNB, and surgical resection from the center and periphery of the tumor. Profiles were generated by reverse phase protein arrays (RPPA) and immunohistochemistry (IHC). Intraoperative ultrasound-guided percutaneous FNAs (4–8 passes) and CNBs (3–4 core biopsies) were obtained from 53 patients with lesions greater than 2 cm prior to surgical resection. FNA specimens, the largest CNB, and portions of the surgical resection from each patient were frozen for RPPA, and the remaining CNB and surgical resection specimens were formalin-fixed and paraffin-embedded for H&E and IHC staining. RPPA investigated levels of 154 proteins (enriched for PI3K/Akt/mTOR pathways) in specimens with greater than 20% tumor content on the corresponding H&E specimen (44 CNBs, 46 resection specimens from the tumor center, and 43 resection specimens from the tumor periphery). RPPA was normalized for FNA and CNB specimens independently. For each specimen an overall PI3K pathway score was calculated by adding normalized values and the specimen was classified as active if the sum was in the top quarter for their sample type.

   Summary of Findings:

   There were no differences in tumor content between resection specimens from the tumor center and periphery; however, CNB specimens had significantly lower tumor content than resection specimens from the periphery (52.7% versus 60.7%, P=0.048). Functional protein profiles of FNA specimens clustered separately but CNB and resection specimens clustered together by patient rather than by procurement methods. Levels of phosphorylated proteins were less correlated among procurement methods than non-phosphorylated proteins, but significant differences were limited to FNA and resection specimens (P=0.04). Correlations between IHC and RPPA staining were noted for pAkt S473 (r=0.71), PTEN (r=0.41), and pS6 S235/236 (r=0.33) among resection specimens from the tumor center and for pAkt S473 in surgical resection specimens from the tumor periphery (r=0.65), and CNB specimens (r=0.52). Expression levels in resection specimens from the tumor periphery and center were significantly correlated for 149 of the 154 proteins evaluated, but levels of caveolin, eEF2, pJNK T183/185, p-mTOR S2448, PRAS40 T246, and pSrc Y527 were not. Importantly, the strength of correlations for RPPA-determined protein levels varied in a protein-specific manner between resection specimens from the center and periphery: ER (r=0.90), PR (r=0.90), HER2 (r=0.79), pAKT S473 (r=0.70), pAkt T308 (r=0.61), PTEN (r=0.72), INPP4B (r=0.77), p4E-BP1 T37/46 (r=0.62), p4E-BP1 S65 (r=0.53), pS6 S235/236 (r=0.42), and pS6 240/244 (r=0.44). Using a false discovery rate (FDR) of 0.05, there was no difference in expression between the periphery and the core for 132 of the 154 markers (86%) and there was 74.4% agreement in PI3K activation status. Similarly, IHC showed modest to strong correlations in p4E-BP1 T70, pAkt S473 pPRAS40 T246, pS6 S240/244, pS6 S235/236, and PTEN expression between specimens from the tumor center and periphery (r=0.81, r=0.588, r=0.82, r=0.55, r=0.48, r=0.82, r=0.55, r=0.48, and r=0.85, respectively).

   Using a FDR of 0.05, resection specimens and CNB had comparable levels of 97 of the 154 markers (67%). Importantly, weak to strong correlations for RPPA levels between resection specimens and CNB were observed for levels of ER (r=0.80), PR (r=0.87), HER2 (r=0.66), pAKT S473 (r=0.60), pAkt T308 (r=0.60), PTEN (r=0.72), INPP4B (r=0.77), p4E-BP1 T37/46 (r=0.35), p4E-BP1 S65 (r-0.32), and pS6 240/244 (r=0.36). Correlations of similar strength were observed for IHC staining between CNB and the resection specimen from the core for pAkt S473, PTEN, p4E-BP1 T70 and pS6 S235/236 (r=0.87, r=0.72, r=0.81, and r-0.38, respectively), and between the CNB and peripheral resection specimen for pAkt S473, PTEN, p4E-BP1 S65, pS6 S240/244, and pS6 S235/236 (r=0.71, r=0.68, r=0.63, r=0.47, and r=0.49, respectively). There was 78.4% agreement in PI3K activation status between CNB and resection specimens from the tumor center. Compared to resection specimens from the core, PI3K activation was higher in CNB as determined by RPPA levels of pAkt S473 (P=0.02), p4E-BP1 S65 (P=0.03), p4E-BP1 S37/46 (P<0.001), pS6 S235/236 (P=0.01) pS6 S240/244 (P<0.001), PTEN (P<0.001), and pMAPK T202/204 (P=0.004), but only p4E-BP1 T37/46 and pS6 240/244 levels were significantly higher in CNB than the resection specimen (P=0.005 and P<0.001, respectively).

   In contrast, of the 154 proteins evaluated, only 49 proteins were significantly correlated between FNA and resection specimens and 68 proteins between FNA and CNB specimens (p<0.05). Modest correlations between FNA and surgical resection or CNB specimens were noted for ER (r=0.606 and r=0.650, respectively) and PR (r=.70 and r=0.64, respectively) and a weak-modest correlation for HER-2 levels (r=0.37 and r=0.57, respectively). IHC staining of pAkt S473, PTEN, and pS6 S235/236 in the resection specimen was weakly correlated with RPPA levels in the FNA specimen (r=0.30, r=0.37 and r=0.35, respectively).

   Biospecimens

   • Tissue - Breast

   Preservative Types

   • Frozen
   • Formalin

   Diagnoses:

   • Neoplastic - Carcinoma

   Platform:

   Analyte	Technology Platform
   Protein	Immunohistochemistry
   Protein	Reverse phase protein microarray

   Pre-analytical Factors:

   Classification	Pre-analytical Factor	Value(s)
   Biospecimen Acquisition	Method of tissue acquisition	Core needle biopsy Fine needle aspiration Surgical resection
   Immunohistochemistry Specific	Targeted peptide/protein	ki-67 PTEN pAkt S473 pS6 S235/236 pS6 S240/244 pPRAS40 T246 p4EBP1 T70
   Immunohistochemistry Specific	Technology platform	RPPA

   View more

2. Study Purpose

   This study investigated the effects of tumor size and grade, specimen cellularity, cold ischemia time, surgical procedure, prior chemotherapy, and patient age on correlated protein expression among FNA, CNB, and surgical resection specimens from the tumor center and periphery. Intraoperative ultrasound-guided percutaneous FNAs (4–8 passes) and CNBs (3–4 core biopsies) were obtained from 53 patients with lesions larger than 2 cm prior to surgical resection. FNA specimens, the largest CNB, and portions of each surgical resection from each patient were frozen for RPPA which evaluated levels of 154 proteins (enriched for PI3K/Akt/mTOR pathways) in specimens with greater than 20% tumor content on the corresponding H&E specimen. RPPA was normalized for FNA and CNB specimens independently. 

   Summary of Findings:

   Specimens exposed to short cold ischemia time (<36.7 min versus >36.7 min) displayed stronger correlations in proteins levels between specimens from the tumor center and periphery (P<0.001), CNB and surgical specimens (P<0.001), and FNA specimens and surgical specimens from the tumor center (P<0.001). Further, specimens resulting from breast conservation surgery had significantly shorter mean cold ischemia times than mastectomy specimens (28.5 min versus 39.1 min, P=0.021), and protein expression was more strongly correlated with that in CNB or FNA among specimens procured by breast conservation than mastectomy and (P<0.005 and P<0.01, respectively). A trend toward stronger correlated protein expression was also observed between surgical breast conservation specimens from the tumor center and periphery.

   Tumors that were smaller than 3.6 cm displayed more strongly correlated protein levels between surgical specimens from the tumor center and periphery (P<0.001), while weaker correlations in expression were observed between FNA and resection specimens (P=0.003). Although none of the proteins were differentially expressed between tumors of different grades or sizes across procurement methods, pGSK3a/b S21/9 expression was higher in grade 3 tumors compared to grade 1 or 2 tumors when specimens from all procurement methods were combined (P=0.0343 and P=0.0001, respectively). Tumor cellularity (20-49% versus more than 50%) significantly affected the expression of 54 proteins in CNB specimens and 22 proteins in surgical specimens from the tumor center, with 7 common among the two procurement methods (Cyclin B1, Beclin, IRS, YB1, p38 MAPK, BID and eIF4G); none of the observed differences in protein levels due to tumor cellularity were explained by tumor cell composition.

   Specimens from patients who received neoadjuvant chemotherapy displayed protein expression levels that were more weakly correlated between surgical specimens from the tumor core and periphery (P<0.0001), CNB and surgical specimens (P<0.0001), FNA and CNB (P<0.0001), and FNA and surgical specimens from the periphery (P=0.0209) compared to specimens from patients who did not receive treatment.

   When compared to younger patients, patients older than 51 years had higher levels of ER in CNB and resection specimens from the center or periphery (P<0.0001, all) and lower levels of TIGAR and c-Kit in CNB specimens (P=0.0008 and P=0.01, respectively).

   Biospecimens

   • Tissue - Breast

   Preservative Types

   • Frozen

   Diagnoses:

   • Neoplastic - Carcinoma

   Platform:

   Analyte	Technology Platform
   Protein	Reverse phase protein microarray

   Pre-analytical Factors:

   Classification	Pre-analytical Factor	Value(s)
   Preaquisition	Patient age	<51 years >51 years
   Preaquisition	Other drugs	Received neoadjuvant chemotherapy No chemotherapy
   Preaquisition	Surgical procedure type	Mastectomy Breast conservation surgery
   Preaquisition	Prognostic factor	Tumor <3.6 cm Tumor >3.6 cm Grade 1 tumor Grade 2 tumor Grade 3 tumor
   Biospecimen Acquisition	Cold ischemia time	8-36.7 min 36.8-157 min
   Biospecimen Acquisition	Method of tissue acquisition	Fine needle aspiration Core needle biopsy Mastectomy Surgical resection Lumpectomy
   Biospecimen Aliquots and Components	Biospecimen heterogeneity	Between core and periphery Differing levels of cellularity

   View more