Abstract PD9-05: The importance of the metastatic biopsy: Clinical and translational relevance in a real world series of patients with metastatic breast cancer

Background

Metastatic breast cancer (MBC) is a heterogeneous disease, whose clinical course and prognosis may be unpredictable, creating significant uncertainty for patients and their families. Heterogeneity is breast cancer subtypes is now well recognized as a potential reason for treatment resistance. Sampling metastatic sites at the point of diagnosis or upon progression, when safe, is recommended to better guide therapy.

Purpose

This study evaluated patients currently undergoing treatment for MBC in the clinic to determine the clinical and translational significance of a metastatic sample.

Methods

Patients currently undergoing treatment for MBC at the Olivia Newton John Centre were identified. Data was collected on patient demographics, clinicopathological information, treatment and duration of response. Translational research tissue was collected, with consent, for DNA and RNA analysis.

Results

Between January 2017 and May 2018 111 patients were identified. The mean age of MBC diagnosis was 60 years (range 30-87), with a mean follow up time of 2.4 years (range 0.8-16). Fifteen patients died during the study period. Sixty-seven (60%) patients were initially treated for early breast cancer (EBC), with a median disease free interval (DFI) 4.7 years. Half (51%) these patients relapsed after five years.

At MBC diagnosis, multiple sites of disease were identified including bone, visceral, brain, nodal and skin/chest wall disease. Bone only disease was common (25%), whereas brain disease was rare overall (9%). Metastatic tissue was collected in 67 (60%) patients, where up to four different sites were biopsied. The most commonly biopsied site was bone (n=21), followed by soft tissue (n=20), chest wall/skin disease (n=12), liver (n=9), lung (n=8) and brain (n=8). Serous disease was collected in 16 patients, including pleural, pericardial, ascitic and cerebrospinal fluid.

Based on the EBC subtype (n=67), 70% had luminal disease, 19% had Her2 positive disease and 7% had TNBC. However, based on a metastatic biopsy (n=67), only 61% of patients had luminal disease, 21% had HER2 positive, and 18% had TNBC. Paired EBC and MBC samples were available in 48 patients, with significant change in breast cancer subtype demonstrated in 12 of these patients (25%). The most common change was a loss in ER staining, which included 6 patients from ER positive, HER2 negative to TNBC and three patients who became ER negative but remained HER2 positive. Molecular profiling was performed thus far on 8 samples at the single cell and bulk level. These results highlight a large level of inter- and intra-tumoral heterogeneity, and may result in a better understanding of the molecular pathways specifically deregulated in patients at the point of progression.

Conclusion

In this single institution series of patients with MBC, over half of the cohort underwent at least one metastatic biopsy. Strikingly, a quarter of patients demonstrated a change in their breast cancer subtype, which directly guided subsequent therapy. Metastatic tissue can provide vital information to inform treatment decisions, which may be guided by translational laboratories having access to fresh tissue at the point of metastatic diagnosis or disease progression.

Citation Format: Yeo B, Molinaro T, Merino D, Berthelet J, Pouliot N, Fang C, Bell C, Anderson R. The importance of the metastatic biopsy: Clinical and translational relevance in a real world series of patients with metastatic breast cancer [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr PD9-05.

Laminin-10 mediates basal and EGF-stimulated motility of human colon carcinoma cells via alpha(3)beta(1) and alpha(6)beta(4) integrins

Signals from the epidermal growth factor (EGF) receptor and integrin-dependent adhesion to laminin contribute to the progression and metastasis of colonic tumors. However, little is know about the mechanisms by which these signals cooperate. Recently, we have reported that the colon cancer cell line LIM1215 secretes and adhere to autocrine laminin-10 via multiple integrin receptors and that EGF stimulates spreading of these cells on the same substrate. In this report, we investigate the effect of EGF and laminin-10 on colon cancer cell migration in vitro. EGF stimulates migration of LIM1215 cells in a wound healing assay. The response to EGF is inhibited by anti-EGF receptor antibody 528, the EGF receptor kinase inhibitor AG-1478, or the MAP kinase kinase inhibitor PD98059 but not the PI3-K inhibitor wortmannin. Using Transwell migration chambers, we demonstrate that laminin-10 but not collagen-I, collagen-IV, or a commercial preparation of human placental laminin is a potent motility factor for LIM1215 cells. The migration response to laminin-10 is increased upon stimulation of the cells with EGF and correlates with the up-regulation of alpha(6)beta(4) integrin expression as measured by analysis of Triton X-100-soluble cellular extracts. The results from integrin inhibition experiments indicate that basal migration on laminin-10 is mediated by alpha(3)beta(1) but not alpha(2)beta(1) nor alpha(6)beta(4) integrins. Alpha(3) blocking antibodies also inhibited EGF-stimulated chemokinetic migration of LIM1215 cells on laminin-10. However, in contrast to unstimulated cells, alpha(6) or beta(4) integrin-blocking antibodies inhibited the migration of EGF-stimulated cells by up to 50%. Taken together, these results support the cooperative role of EGF receptor and laminin-10 on colon cancer cell motility and suggest a critical role for both the alpha(3)beta(1) and the alpha(6)beta(4) integrins in this process.

Colon cancer cells adhesion and spreading on autocrine laminin-10 is mediated by multiple integrin receptors and modulated by EGF receptor stimulation

Epidermal growth factor (EGF) receptor ligands such as EGF and transforming growth factor-alpha (TGF-alpha) play an important role in controlling the proliferation, survival, morphology, and motility of colonic epithelial cells. There is also increasing evidence that growth factors and extracellular matrix (ECM) proteins cooperate to regulate these cellular processes. We have reported previously that autocrine TGF-alpha and an unidentified ECM protein in the serum-free conditioned medium of the human colon carcinoma cell line LIM1215 synergize to induce spreading of these cells in low-density cultures. We have now purified the ECM protein secreted by LIM1215 cells and show that it synergizes with EGF to induce spreading of LIM1215 cells and other human cell lines from the colon and other tissues. The purified ECM migrated as a single protein band with an apparent molecular mass of approximately 800 kDa on SDS-PAGE under nonreducing conditions and, under reducing conditions, as three protein bands of approximately 360, 210, and 200 kDa. Immunoblotting experiments and mass spectrometry analysis of tryptic digests on the purified protein identified the 360-, 210-, and 200-kDa protein bands as laminin alpha5, beta1, and gamma1 chains, respectively, indicating that LIM1215 cells secrete laminin-10 (alpha5 beta1 gamma1). In serum-free medium, LIM1215 cells adhere to laminin-10 primarily via alpha2 beta1 and alpha3 beta1 integrin receptors. EGF-induced spreading of LIM1215 cells on laminin-10 is partially inhibited by pretreatment of the cells with blocking antibodies directed against integrin alpha3 or beta1 but not alpha2, alpha6, or beta4 subunits. Spreading is almost completely inhibited by blocking alpha3 + alpha2, alpha3 + alpha6, or beta1 + beta4 integrin chains and results in cell death. Increased spreading in the presence of EGF correlates with up-regulation of alpha6 beta4 integrins in these cells after exposure to EGF. These results indicate that colon cancer cells attach and spread on laminin-10 via multiple integrin receptors and suggest a critical role for alpha3 beta1 integrins in the spreading response. Together, our results support the concept that the adhesive properties of colon cancer cells are modulated by autocrine production of TGF-alpha and laminin-10 and autocrine induction of appropriate integrins.

Multiple autocrine factors including an extracellular matrix protein are required for the proliferation and spreading of human colon carcinoma cells in vitro

The human colon carcinoma cell line LIM1215 proliferates and changes morphology (spread) in a cell density-dependent manner in response to epidermal growth factor (EGF). At high density, production of autocrine transforming growth factor-alpha enables the cells to proliferate and spread in the absence of exogenous EGF or serum. At low cell density (< 1 x 10(4)/cm2) EGF alone fails to elicit a mitogenic or morphological response and requires the presence of conditioned medium (derived from high cell density serum-free culture of the same cells) to exert its effects. This synergy between EGF and LIM1215 conditioned medium was investigated further. Using a low cell density assay and fractionated LIM1215 conditioned medium, we show that EGF-mediated mitogenic and morphological responses are separable. These responses are dependent on the synergistic action of a low molecular weight autocrine survival factor and an extracellular matrix-like spreading factor(s) secreted into the culture medium respectively. We find that under low cell density, serum-free conditions, EGF alone is insufficient to rescue LIM1215 from rapid apoptotic death. Catalase or LIM1215 autocrine survival factor prevent the death of LIM1215 cells and restore their proliferative (but not morphological) response to EGF, suggesting that cell death under these conditions may be the result of oxidative stress. Combination of EGF, partially purified autocrine survival and spreading factors induced proliferation and spreading of low density LIM1215 cells similar to that observed with EGF and unfractionated conditioned medium. GRGDS peptides strongly inhibited the spreading of LIM1215 cells in the presence of EGF and the partially purified autocrine spreading factor, demonstrating that integrin receptors are involved in the spreading process. Comparison of the spreading response of LIM1215 and Colo 526 cells on ASF and various adhesion proteins indicate that ASF is not collagen-I, collagen-IV, fibronectin or vitronectin. Taken together, these results support the concept that the autonomous growth of colon carcinoma cells in vitro is dependent on the synergistic interaction between several autocrine systems.

Activation of the Ras/mitogen-activated protein kinase pathway by kinase-defective epidermal growth factor receptors results in cell survival but not proliferation

Signalling by the epidermal growth factor (EGF) receptor (EGFR) has been studied intensively, but for most cell types the analysis is complicated by the fact that EGFR not only homodimerizes but can also form heterodimers with other EGFR family members. Heterodimerization is a particular problem in the study of EGFR mutants, where the true phenotype of the mutants is confounded by the contribution of the heterodimer partner to signal transduction. We have made use of the murine hemopoietic cell line BaF/3, which does not express EGFR family members, to express wild-type (WT) EGFR, three kinase-defective EGFR mutants (V741G, Y740F, and K721R), or a C-terminally truncated EGFR (CT957) and have measured their responses to EGF. We found that under the appropriate conditions EGF can stimulate cell proliferation of BaF/3 cells expressing WT or CT957 EGFRs but not that of cells expressing the kinase-defective mutants. However, EGF promotes the survival of BaF/3 cells expressing either of the kinase-defective receptors (V741G and Y740F), indicating that these receptors can still transmit a survival signal. Analysis of the early signalling events by the WT, V741G, and Y740F mutant EGF receptors indicated that EGF stimulates comparable levels of Shc phosphorylation, Shc-GRB-2 association, and activation of Ras, B-Raf, and Erk-1. Blocking the mitogen-activated protein kinase (MAPK) signalling pathway with the specific inhibitor PD98059 abrogates completely the EGF-dependent survival of cells expressing the kinase-defective EGFR mutants but has no effect on the EGF-dependent proliferation mediated by WT and CT957 EGFRs. Similarly, the Src family kinase inhibitor PP1 abrogates EGF-dependent survival without affecting proliferation. However blocking phosphatidylinositol-3-kinase or JAK-2 kinase with specific inhibitors does arrest growth factor-dependent cell proliferation. Thus, EGFR-mediated mitogenic signalling in BaF/3 cells requires an intact EGFR tyrosine kinase activity and appears to depend on the activation of both the JAK-2 and PI-3 kinase pathways. Activation of the Src family of kinases or of the Ras/MAPK pathway can, however, be initiated by a kinase-impaired EGFR and is linked to survival.

Natural killer and lectin-dependent cytotoxic activities of Kurloff cells: target cell selectivity, conjugate formation, and Ca++ dependency

Kurloff cells may represent a major component of NK cell activity in the guinea pig. We have pursued to characterize the mechanism of their action. Using murine target cells, we found Kurloff cell cytotoxicity to be selective for the NK-sensitive YAC-1 target cell, with minimal activity against the NK-resistant P815 target cell. In the presence of PHA, but not ConA, cytotoxicity was markedly augmented against both YAC-1 and P815. While effector-target conjugate formation was observed with YAC-1 cells but not P815 cells in control cultures, it was augmented with both target cell types in cultures with PHA. Pretreatment alone with PHA was ineffective, however. NK cell activity of Kurloff cells was dependent on extracellular Ca++ and entry of Ca++ into the effector cells, as demonstrated by abrogation of cytotoxicity when extracellular Ca++ was chelated with EDTA or EGTA, or following treatment with the Ca++ channel blockers verapamil and diltiazem. Furthermore, inhibition of PKC by H7 resulted in significant reduction of Kurloff cell-mediated NK activity, while pretreatment of effector cells with the PKC activator TPA enhanced NK activity. Kurloff cells could also be stimulated to produce serine esterases by contact with target cells or treatment with phorbol ester and ionophore. Finally, a majority of Kurloff cells, identified by the monoclonal antibody 14D1, reacted with the human NK cell marker CD56. Taken together, these data suggest that Kurloff cells have NK-like characteristics and activity, with target cell selectivity, and that their lytic mechanisms involve influx of extracellular Ca++, PKC activation and serine esterase production.

Guinea pig Kurloff (NK-like) cells mediate TNF-dependent cytotoxic activity: analogy with NC effector cells

Kurloff cells are mononuclear cells possessing a large cytoplasmic inclusion body specific to the guinea pig. In this report, we present strong evidence that Kurloff cells can mediate NC activity against tumor cells in addition to their previously reported NK activity. Using an 18 h 51Cr-release assay we have shown that Kurloff cells were highly effective in killing the TNF-sensitive WEHI 164 target cell line. Lower but significant cytotoxic activity was also observed after only 4 h. However, our results suggest a different mechanism of lysis in the 4 h and 18 h assay. Lysis of WEHI 164 target cells by Kurloff cells in the 4 h assay could be strongly increased in the presence of TPA alone or in combination with ionomycin whereas ionomycin alone was uneffective. In contrast, stimulation of Kurloff cells for 18 h with ionomycin alone or in combination with TPA could induce the release of TNF-like factor(s) as observed by the TNF bioassay using L-929 TNF-sensitive target cells. Release of TNF-like factor(s) could also be induced by stimulation with WEHI 164 target cells. Supernatants of Kurloff cells stimulated for 18 h with TPA + ionomycin were also highly cytotoxic against WEHI 164 target cells, but not against the TNF-resistant P815 target cell line. Pretreatment of these supernatants with antimurine TNF alpha antibodies could almost completely inhibit their cytotoxic activity against WEHI 164 target cells. In contrast, supernatants of Kurloff cells stimulated for only 4 h did not show any TNF-like activity against the L-929 target cell line and were not cytotoxic against WEHI 164 target cells even after 18 h. Taken together, these results suggest that Kurloff cells can mediate NC activity against tumor cells in addition to their previously reported NK activity. By using multiple lytic pathways, these cells may play a crucial role in anti-tumor surveillance and defenses.

Maximum likelihood positioning in the scintillation camera using depth of interaction

Specific effects of the depth of interaction (DOI) on the photomultiplier (PM) response in an Auger gamma camera were quantified. The method was implemented and tested on a Monte Carlo simulator with special care to the noise modeling. Two models were developed, one considering only the geometric aspects of the camera and used for comparison, and one describing a more realistic camera environment. In a typical camera configuration and 140-keV photons, the DOI alone can account for a 6.4-mm discrepancy in position and 12% in energy between two scintillations. Variation of the DOI can still bring additional distortions when photons do not enter the crystal perpendicularly such as in slant hole, cone beam and other focusing collimators. With a 0.95-cm crystal and a 30 degrees slant angle, the obliquity factor can be responsible for a 5.5-mm variation in the event position. Results indicate that both geometrical and stochastic effects of the DOI are definitely reducing the camera performances and should be included in the image formation process.

Statistical and physical content of low-energy photons in holospectral imaging

The utility of low-energy photons can be investigated using statistical and physical models. Holospectral imaging exploits the correlation between energy frames to optimize primary photon representation. A relationship between energy frames can be established by the physical description of low-energy photons (or events). It is shown that those events not only build an image with fortuitous correlation with the on-peak image but, in some cases, they are spatially indistinguishable from primary photons. Similarly, it is possible to show that an important proportion of the events detected at the primary energy are in fact distributed in more than one pixel and, consequently, are misplaced and should be rejected from the imaging process. The authors introduce the rationale for statistical and physical approaches and explain the various categories of photons considered in this study. Monte Carlo simulations were used to establish the rate of occurrence of each photon path.