Characterisation of the immune microenvironment of primary breast cancer and brain metastasis reveals depleted T-cell response associated to ARG2 expression

BACKGROUND

Immune checkpoint inhibition is an established treatment in programmed death-ligand 1 (PD-L1)-positive metastatic triple-negative (TN) breast cancer (BC). However, the immune landscape of breast cancer brain metastasis (BCBM) remains poorly defined.

MATERIALS AND METHODS

The tumour-infiltrating lymphocytes (TILs) and the messenger RNA (mRNA) levels of 770 immune-related genes (NanoString™, nCounter™ Immuno-oncology IO360) were assessed in primary BCs and BCBMs. The prognostic role of ARG2 transcripts and protein expression in primary BCs and its association with outcome was determined.

RESULTS

There was a significant reduction of TILs in the BCBMs in comparison to primary BCs. 11.5% of BCs presented a high immune infiltrate (hot), 46.2% were altered (immunosuppressed/excluded) and 34.6% were cold (no/low immune infiltrate). 3.8% of BCBMs were hot, 23.1% altered and 73.1% cold. One hundred and twelve immune-related genes including PD-L1 and CTLA4 were decreased in BCBM compared to the primary BCs (false discovery rate <0.01, log2 fold-change >1.5). These genes are involved in matrix remodelling and metastasis, cytokine-chemokine signalling, lymphoid compartment, antigen presentation and immune cell adhesion and migration. Immuno-modulators such as PD-L1 (CD274), CTLA4, TIGIT and CD276 (B7H3) were decreased in BCBMs. However, PD-L1 and CTLA4 expression was significantly higher in TN BCBMs (P = 0.01), with CTLA4 expression also high in human epidermal growth factor receptor 2-positive (P < 0.01) compared to estrogen receptor-positive BCBMs. ARG2 was one of four genes up-regulated in BCBMs. High ARG2 mRNA expression in primary BCs was associated with worse distant metastasis-free survival (P = 0.038), while ARG2 protein expression was associated with worse breast-brain metastasis-free (P = 0.027) and overall survival (P = 0.019). High transcript levels of ARG2 correlated to low levels of cytotoxic and T cells in both BC and BCBM (P < 0.01).

CONCLUSION

This study highlights the immunological differences between primary BCs and BCBMs and the potential importance of ARG2 expression in T-cell depletion and clinical outcome.

Abstract P6-05-02: Identification of microRNAs differentially expressed in brain metastasis secondary to breast cancer

Background: Despite sequential improvements in the adjuvant treatment of breast cancer (BC), recurrence and metastasis remains a major clinical problem and in particular, brain metastasis (BCBM). A number of microRNAs (miRNAs) have been linked to the metastatic process in BC, but to date there is limited work on the microRNAs involved in BCBM. The current study aim to identify differentially expressed miRNAs within primary breast cancer who did not recur (BCNR) versus primary BC cases which did recur (BCR) and their matched BCBM cases.

Methods: Formalin-fixed paraffin-embedded (FFPE) material was collected of 12 primary BCNRs from the Liverpool tissue bank and of 40 paired primary BCR samples and their matched BCBM from the Walton Research Tissue Bank and RCSI National Breast Cancer Bioresource. miRNA was extracted (Qiagen miRNeasy FFPE kit) and profiled using the NanoString™ nCounter™ miRNA Expression Assay (Human v3 miRNA). The differentially expressed miRNAs between BCNR versus BCR and BCR versus their matched BCBM were identified by significance of microarray analysis (SAM) on the MeV4.9 software. Pathway analysis was performed using the DIANA-mirPath v3.0 software and the Ingenuity Pathway Analysis (IPA) to identify a network of genes/pathways regulated by the differentially expressed miRNAs.

Results: 12 BCNR and 30 matched pairs of BCR and BCBM passed the quality control and normalisation processes. Principal component analysis (PCA) performed on 166 miRNAs after QC/normalisation clearly distinguishes the BCNR and the primary BCR from the matched BCBM cases, whereas SAM revealed 58 differentially expressed miRNAs with a 10% FDR (false discovery rate) and an absolute log2 fold-change (FC) &gt;1 between BCNR and BCR and 11 between the matched BCs and BCBMs. Pathway clustering revealed that these differentially expressed miRNAs (10% FDR, log2FC&gt;1) within both BCNR vs BCR and BCR vs BCBM cohorts are highly enriched for genes involved in extracellular matrix (ECM)-receptor interactions, proteoglycans, adherens junctions, TGF-β, p53 and Hippo signalling. IPA identified a network of genes, implicated in the processes of breast cancer invasion and metastasis, regulated by the identified miRNAs, such as, TWIST, MET, TP53, MYC, EZH2, ZEB1, TAGLN and BIRC5. Four of the significantly differentially expressed miRNAs, hsa-miR-132-3p, hsa-miR-199a-5p, hsa-miR-150-5p and hsa-miR-155-5p were present within both cohorts (BCNR vs BCR and BCR vs BCBM) and regulate genes involved in Hippo and TGF-β signalling (DIANA-mirPath v3.0 analysis: p=5.23x10-08 and p=2.67x10-07 respectively).

Conclusion: The current study, utilising a large cohort of paired BCR and BCBM cases, provides novel insight into the molecular mechanisms and role of miRNAs in BCBM. Four miRNAs (hsa-miR-132-3p, hsa-miR-199a-5p, hsa-miR-150-5p and hsa-miR-155-5p) in particular could be potentially used to identify patients with increased risk of developing brain metastasis and help facilitate the development of specific treatments for BCBM, which to date have proved elusive. The miRNAs identified require further exploration as potential biomarkers as well as novel therapeutic targets.

Citation Format: Giannoudis A, Clarke K, Zakaria R, Vareslija D, Farahani M, Rainbow L, Platt-Higgins A, Ruthven S, Brougham K, Rudland PS, Jenkinson MD, Young L, Falciani F, Palmieri C. Identification of microRNAs differentially expressed in brain metastasis secondary to 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 P6-05-02.

Abstract P2-03-04: Application of digital-PCR technology to determine c-MET copy number variation in paired primary breast cancer and brain metastases

INTRODUCTION:

c-MET amplification/overexpression has been associated with treatment failure and progression in many cancers, including breast cancer (BC). c-MET showed amplification by fluorescent in situ hybridization (FISH) in 27% of trastuzumab-treated HER2-positive patients. These patients had a high trastuzumab failure rate and a shorter time to progression. Up to 50% of patients with metastatic HER2-positive disease will develop brain metastases (BM) during their disease course and in approximately one third, brain is the first site of progression. Amplification/copy number variations (CNVs) are mainly assessed by FISH whereas overexpression is assessed by immunohistochemistry (IHC). We present a PCR-based assay (digital-PCR) able to determine CNV in c-MET and HER2 in a cohort of patients with metastatic BC to the brain and demonstrate the correlation of CNV to protein expression.

METHODS:

DNA was isolated from paraffin-embedded tissues of 23 paired primary BC-BM cases. CNV was analysed by the QuantStudioTM 3D-Digital-PCR (QS3D) and real-time qPCR (both from ThermoFisher Scientific). The breast MCF7, T47D, BT474, AU565, SKBR3 and the gastric MKN45 cell lines were used as controls for the HER2 and c-MET CNV assays. Copy number per diploid genome was calculated using the absolute quantification number of FAM-labelled target and VIC-labelled RNaseP reference multiplied by 2. Cases with ≤2 copies are classified as normal whereas cases with &gt;2 were classified as amplified. The HER2 positivity of the primary BC cases was routinely assessed by IHC. The c-MET protein expression was assessed by IHC using the c-MET(3D4) monoclonal antibody (ThermoFisher Scientific).

RESULTS:

CNV in c-MET by QS3D digital-PCR was detected in 69.6% of primary BC (ER-/HER2+:2, ER+/HER2+:5, ER+/HER2-:8, Triple-negatives:5, unknown:3) as well as 69.6% of BM, whereas HER2 CNV was observed in 39.1% primary BC and 52.2% BM. In the HER2-positive cases, the prevalence of HER2 CNV was 100% in both primary BC and BM. Within these cases, c-MET CNV was 85.7% in the primary BC and 71.43% in BM. CNVs in both genes were observed in 30.4 % of all primary and 39.1% of BM. The CNV data are presented in Table 1.

CNV data for c-MET and HER2 between primary breast cancer (BC) and corresponding brain metastatic (BM) casesc-MET CNVHER2 CNVBCBMBCBM&gt;2: 16&gt;2: 12≤2: 4&gt;2: 9&gt;2: 8≤2: 1≤2: 7&gt;2: 4≤2: 3≤2: 14&gt;2: 4≤2: 10

There was a high concordance between the QS3D and qPCR data with Pearson's R=0.74 (p&lt;0.00001).

A significant correlation between HER2 protein expression and CNV was observed (Fisher's exact test p=0.0005). Data will be presented on c-MET protein expression in the pair samples.

CONCLUSIONS:

The prevalence of CNV is much higher than that reported by immunohistochemistry and FISH in the literature to date, possibly due to the sensitivity of the digital-PCR technology. The high level of c-MET CNV in primary and metastatic BC, and the concurrent CNV in both genes warrants further investigation. It also highlights the potential to use c-MET directed therapy particularly in HER2+ BC and reinforces the potential importance of precise detection methods in both the primary and metastatic setting. Analysis of a larger series is currently on-going.

Citation Format: Giannoudis A, Zakaria R, Platt-Higgins A, Syed KAR, Ashton K, Dawson T, Rudland PS, Holcombe C, Jenkinson MD, Palmieri C. Application of digital-PCR technology to determine c-MET copy number variation in paired primary breast cancer and brain metastases [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-03-04.

Significance of the metastasis-inducing protein AGR2 for outcome in hormonally treated breast cancer patients

The anterior gradient protein-2 (AGR2) is inducible by oestrogen and itself can induce metastasis in a rat model for breast cancer. Here, a rabbit antibody to recombinant human AGR2 was used to assess its prognostic significance in a retrospective cohort of 351 breast cancer patients treated by adjuvant hormonal therapy. The antibody stains 66% of breast carcinomas to varying degrees. The percentage of positive carcinoma cells in tumours directly correlates with the level of AGR2 mRNA (Spearman's rank correlation, P = 0.0007) and protein (linear regression analysis r2 = 0.95, P = 0.0002). There is a significant association of staining of carcinomas for AGR2 with oestrogen receptor alpha (ERalpha) staining and with low histological grade (both Fisher's Exact test P<0.0001). In the ERalpha-positive cases, but not the ERalpha-negative cases, when subdivided into the separate staining classes for AGR2, there is a significantly progressive decrease in patient survival with increased staining (log rank test, P = 0.006). The significant association of staining for AGR2 with patient death over a 10-year period (log rank test P = 0.007, hazard ratio = 3) only becomes significant at 6 years of follow-up. This may be due to the cessation of adjuvant hormonal therapy at an earlier time, resulting in adverse re-expression of the metastasis-inducing protein AGR2.

Expression of calcium-binding protein S100A2 in breast lesions

A suppression subtraction cDNA library representing mRNAs expressed at a higher level in a benign breast tumour-derived cell line relative to the malignant MCF-7A cell line contained cDNAs corresponding to mRNAs for plasminogen activator inhibitor I, annexin VIII and the EF-hand protein S100A2. S100A2 protein has previously been shown to be expressed in normal human breast epithelium, but not in human breast carcinoma cell lines. Using a PCR-based assay and in situ hybridization on histological sections of human breast specimens, the mRNA for S100A2 was shown to be present in all benign breast lesions examined as well as in normal epithelium. S100A2 mRNA was detectable in 37% of specimens of carcinoma in situ, but in less than 15% of carcinoma specimens. The results suggest that the loss of S100A2 is associated with the development of malignant cells and is not associated with early tumour development.

Variant estrogen receptor alpha mRNAs in human breast cancer specimens

A panel of human breast cancer specimens was examined for single base change mutations by DNA sequencing and for larger deletions using a PCR-based assay. In the cancer specimens examined, no sequencing variants were detected other than a previously characterized polymorphism. Although most of the specimens contained estrogen receptor (ER) variants at a low level, 2 of 118 specimens exhibited variants which, after amplification, constituted most of the amplified ER cDNA. One specimen contained a single variant form, and there was little evidence of the wild-type ER mRNA by PCR, Northern blotting or immunocytochemistry. The second specimen, despite the presence of a normal-sized mRNA by Northern blotting and normal immunocytochemical staining for ER, contained at least 5 different variant forms as well as the wild-type ER. All but 1 of the variant forms were processing variants, and 3 of these processing variants have not been described before. One variant, although lacking exons 2-4, has break points in exons 1 and 5 that do not correspond to intron-exon boundaries. This variant might reflect more widespread damage to the genome in this breast cancer specimen. The low level of occurrence of variants suggests that ER variant forms, at least in the coding region, do not contribute generally to the progression of breast cancer.

Variant estrogen receptor alpha mRNAs in human breast cancer specimens

A panel of human breast cancer specimens was examined for single base change mutations by DNA sequencing and for larger deletions using a PCR-based assay. In the cancer specimens examined, no sequencing variants were detected other than a previously characterized polymorphism. Although most of the specimens contained estrogen receptor (ER) variants at a low level, 2 of 118 specimens exhibited variants which, after amplification, constituted most of the amplified ER cDNA. One specimen contained a single variant form, and there was little evidence of the wild-type ER mRNA by PCR, Northern blotting or immunocytochemistry. The second specimen, despite the presence of a normal-sized mRNA by Northern blotting and normal immunocytochemical staining for ER, contained at least 5 different variant forms as well as the wild-type ER. All but 1 of the variant forms were processing variants, and 3 of these processing variants have not been described before. One variant, although lacking exons 2-4, has break points in exons 1 and 5 that do not correspond to intron-exon boundaries. This variant might reflect more widespread damage to the genome in this breast cancer specimen. The low level of occurrence of variants suggests that ER variant forms, at least in the coding region, do not contribute generally to the progression of breast cancer.

Prognostic significance of the metastasis-inducing protein S100A4 (p9Ka) in human breast cancer

The calcium-binding protein S100A4 is capable of inducing metastasis in rodent models for breast cancer. We now show that rabbit antibodies to recombinant rat S100A4 recognize specifically human S100A4 using Western blotting techniques and use them to assess the prognostic significance of S100A4 in primary tumors from a group of 349 patients treated between 1976 and 1982 for stage I and stage II breast cancer. The antibody stains normal breast tissue heterogeneously, but stains positively 41% of the carcinomas, leaving the remaining 59% as negatively stained. In addition to the carcinoma cells, some host stromal cells and lymphocytes are also stained, but these have been discounted in subsequent analyses. There is an association of staining of carcinomas for S100A4 with some tumor variables considered to be associated with poor prognosis for patients: tumor present in axillary lymph nodes (borderline P = 0.058), staining for c-erbB-3 (P = 0.002), cathepsin D (P = 0.024), and c-erbB-2 (P = 0.048). The association of staining for S100A4 with patient survival has been evaluated using life tables and analyzed using generalized Wilcoxon statistics. Eighty percent of the S100A4-negative patients but only 11% of the S100A4-positive patients are alive after 19 years of follow-up, and this association is highly significant (P < 0.0001); the former have a median survival of >228 months and the latter 47 months. The other tumor variables that show significant association with survival time are nodal status (P < 0.0001), tumor size (P = 0.0035), histological grade (P = 0.013), staining for c-erbB-2 (P = 0.0015), estrogen receptor (P = 0.028), and p53 (P = 0.032). Analysis of the association of patients with carcinomas staining for S100A4 and their survival in subgroups defined by these other tumor variables shows that in each subgroup, staining for S100A4 is associated with poorer survival. Patients whose tumors stain for S100A4 and possess involved lymph nodes (P < 0.0001), which are fixed to the chest wall (P = 0.015) or which stain for c-erbB-2 (P = 0.050), show a significant reduction in survival times over those with only S100A4-staining tumors. Patients with involved lymph nodes, or staining for c-erbB-2 in the S100A4-negative group fail to show any significant reduction in survival times. Multivariate regression analysis for 137 patients shows that staining for S100A4 is most highly correlated with patient deaths (P < 0.0001), but involved lymph nodes (P = 0.001), fixed tumors (P = 0.0002), and high histological grade (P = 0.022) are also significant independent prognostic variables. These results suggest that in this group of patients, the metastasis-inducing protein S100A4 is most tightly correlated with patient demise.

Human S100A4 (p9Ka) induces the metastatic phenotype upon benign tumour cells

The rodent S100-related calcium-binding protein, S100A4 induces metastasis in non-metastatic rat and mouse benign mammary cells and co-operates with benign-tumour-inducing changes in two transgenic mouse models, to yield metastatic mammary tumours. Co-transfection of the human gene for S100A4 with pSV2neo into the benign rat mammary cell line, Rama 37, yielded cells which expressed a low level of the endogenous S100A4 mRNA, and either high or undetectable levels of human S100A4 mRNA. The cells which expressed a high level of human S100A4 mRNA induced metastasis in the benign rat mammary cell line Rama 37 in an in vivo assay, whereas the cells which expressed an undetectable level of human S100A4 did not induce any detectable metastases. The primary tumours arising from the S100A4-expressing cells contained high levels of immunocytochemically-detected S100A4 and this high level of S100A4 and the metastatic potential were maintained when cells from a metastasis were re-injected into syngeneic rats. The results show that the human S100A4 possesses metastasis-inducing capabilities.

Cytoplasmic staining of c-erbB-2 is not associated with the presence of detectable c-erbB-2 mRNA in breast cancer specimens

The cell-surface receptor tyrosine kinase protein c-erbB-2 is immunocytochemically detected as membrane staining on the surface of cancer cells in 20-30% of cases of breast cancer, and its presence has been associated with poor prognosis for the patient. However, there have been numerous reports of immunocytochemical staining for c-erbB-2 solely in the cytoplasm of some normal and tumour specimens with frequently used anti-sera, and the presence of such staining has been difficult to interpret. It is not known for certain that cytoplasmic c-erbB-2 staining is an artefact of the immunocytochemical procedures used. Thus, mRNA for c-erbB-2 has been quantified in tumours exhibiting only cytoplasmic staining or varying levels of membrane staining using a sensitive, competitive PCR method. Whereas abundant levels of c-erbB-2 mRNA are found in tumours exhibiting membrane staining for c-erbB-2 and these levels correlate with the percentage of tumour cells showing membranous staining for c-erbB-2, the level of c-erbB-2 mRNA in tumours displaying only cytoplasmic staining is no higher than in c-erbB-2-negative specimens.

Immunocytochemical distribution of the calcium-binding protein p9Ka in normal rat tissues: variation in the cellular location in different tissues

The family of S-100-related proteins consists of a number of small potential calcium-binding proteins of unknown function. Elevated expression of one of these proteins, p9Ka, or of its mRNA, correlates with the metastatic potential of cultured mammary epithelial cells from rat or mouse. Over-expression of p9Ka by transfection of benign rat mammary epithelial tumor cells with the gene for p9Ka induces the metastatic phenotype. At present there is little information on the occurrence of p9Ka in normal rat tissues. A specific antiserum immunocytochemically detects p9Ka intracellularly in most normal adult rat tissues studied, including smooth muscle, brown adipose tissue, and liver. In other tissues, p9Ka is localized specifically to some absorptive and keratinized epithelia, the acid-secreting parietal cells of the stomach, the neuronal cells within plexuses of the autonomic nervous system, and a proportion of cells of the immune system in spleen, lymph nodes, bone marrow, and blood. p9Ka is found widely in both arteries and veins, particularly in the smooth muscle and in the endothelium of smaller veins. In mammary gland, the pattern of staining suggests that p9Ka is extracellularly located in a region surrounding the ducts.

The long term prognostic significance of c-erbB-2 in primary breast cancer

The expression of the c-erbB-2 oncogene has been evaluated using an immunohistochemical technique with the 21N polyclonal antibody in paraffin embedded tissue from 465 patients treated between the years 1975-1981 for Stage I and II breast cancer. One hundred and four (22%) patients exhibited positive staining. This was not associated with any other variables. Expression of the oncogene was associated with significantly poorer survival which was independent of other tumour variables.