Abstract 1725: Tissue-based biomarker analysis of DLBCL using multiplex immunofluorescence AQUA (Automated Quantitative Analysis) algorithms

Purpose: Deeper understanding of immune landscape of DLBCL tumor microenvironment is critical for exploration and development of next generation immunotherapies. Although conventional immunohistochemical (IHC) analysis can provide information on the immune landscape or target expression, it is generally limited to single marker analysis. Advantages of multiplex fluorescence IHC (mFIHC) include the ability to assess different cell types, their spatial relationship as well as variable antigen expression on tumor cells. As such, we developed quantitative mFIHC panels using AQUA® (Automated Quantitative Analysis) technology to evaluate T and B cell populations and their functional status to generate detailed spatial information of immune checkpoint (IC) markers. mFIHC combined with hypothesis driven spatial profiling algorithms (e.g., AQUA Technology) were found to provide the most powerful predictors of immunotherapies in a systematic meta-analyses of over 8000 patients treated with PD1/L1 pathway blockers (Lu et al., JAMA Oncol 2019). Implementation of mFIHC coupled with robust image analysis may provide great insight into DLBCL immune surveillance, mechanism of resistance and patient stratification.

Study Design: We designed two novel mFIHC assays to (1) characterize various B cell populations (CD19, CD20, CD22, PAX5, CD3, TIM3), and (2) evaluate T cell functional status (CD8, Granzyme B, PD1, PDL1, TIM3, Tumor). We successfully validated clinical grade mFIHC assays using automated staining (Leica Bond RX), imaging (Vectra Polaris) and analyses (AQUA Technology) workflow.

Results: Sensitivity, accuracy and specificity were confirmed for all mFIHC assays on known positive and negative controls. Excellent reproducibility (less than 35% CV) and precision were observed across instruments, operators and independent experiments for all markers. Between these two panels, over 200 unique parameters were evaluated. The prevalence of CD19, CD20, CD22, and PAX5 ranged from 30% to >90% positive for the DLBCL samples tested and were overall highly concordant with each other.

Conclusion: The validated mFIHC assays examine B cell antigen expression in DLBCL and their interaction with CD3+ T cells and further characterize CD8+ T cells and immune checkpoint expression. These panels may be used to further understand the complex immune cell and tumor cell spatial biology in the context of clinical trials.

Citation Format: James Paul Santos, Jacob Levy, Kristen Ruma, David Yang, Steve Woolfenden, Jimmie Lim, Xun Li, Ariana Valencia, James Deeds, Ramu Thiruvamoor, Xin Li, Emmanuel Pacia, Margaret McLaughlin, Thai Tran, Sarah Choi, Jennifer Bordeaux. Tissue-based biomarker analysis of DLBCL using multiplex immunofluorescence AQUA (Automated Quantitative Analysis) algorithms [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1725.

Abstract 310: Immunophenotyping tumors by novel multiplex immunofluorescence and AQUA (Automated Quantitative Analyses) algorithms to guide development of immunotherapies

Purpose: Deeper understanding of immune landscape of the tumor microenvironment is critical for exploring and development of next generation immunotherapies. Multiplex fluorescence immunohistochemistry (mFIHC) combined with hypothesis driven spatial profiling algorithms (e.g., AQUA Technology) was found to provide the most powerful predictors of immunotherapies in a systematic meta-analyses of over 8000 patients treated with PD1/L1 pathway blockers (Lu et al., JAMA Oncol 2019).

Study Design: To guide clinical development of next generation and/or combination immunotherapies, we built four novel mFIHC assays to simultaneously explore presence of major immune cell lineages, their spatial relationships and functional attributes. The first assay incorporates antibodies to identify major immune cells lineages (e.g., T-, B-, NK-/T- and subtypes of myeloid cells), the second assay is designed to understand T-cell proliferation (via co-expression of Ki67) or suppression (via co-expression of FoxP3), the third method is focused on quantifying the expression of immune checkpoint inhibitors (e.g., LAG3, PD1 and PD-L1), while, the final assay is designed to explore the IFNγ mediated adaptive resistance and immunosuppression (via co-expression of HLA-DR and IDO1). We will describe the successful development of these clinical grade mFIHC assays that utilize automated staining (Leica Bond RX), imaging (Vectra Polaris) and analyses (AQUA® Technology) workflows on tissue microarrays (TMAs) representing multiple regions of interest including tumor, tumor margin, tumor-associated fibrosis, various adjacent normal tissues, lymphoid structures and/or associated inflammation, and metastatic lesions collected from 100 colorectal, and pancreatic cancer patients.

Results: Sensitivity, accuracy and specificity were confirmed for all mFIHC assays on known positive and negative controls. Excellent reproducibility (less than 35% CV) and precision were observed across instruments, operators and independent experiments for all markers. TMA work is underway and the latest results will be presented.

Conclusion: The validated mFIHC assays helped classify unique immune phenotypes and their interactions. These assays, combined with comprehensive selection of tissues from the tumor microenvironment across multiple cancer subtypes, are expected to enable clinical development of next generation cancer therapies.

Citation Format: Ju Young Kim, John Hoerter, Nathan Roscoe, John Fathman, James Santos, Lori Iaconis, Justin Santos, Minhua Qiu, Jacob Levy, Valerie Tolley, Emmanuel Pacia, Tom Carolan, Xun Li, Patrizia Barzaghi-Rinaudo, Brian Schwartz, Robbin Newlin, Karyn Colman, Andy Schumacher, Deborah Knee, Steven Bender, Margaret McLaughlin, Alexander Savchenko, Thai Tran, Shabnam Tangri, Naveen Dakappagari, Jennifer Bordeaux. Immunophenotyping tumors by novel multiplex immunofluorescence and AQUA (Automated Quantitative Analyses) algorithms to guide development of immunotherapies [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 310.

Abstract P1-09-08: Development and characterization of PTEN IHC assay for testing breast cancer patients specimens

Background: Phosphatase and tensin homolog (PTEN) is a tumor suppressor gene that is a major negative regulator of the Phosphatidylinositol 3-kinase (PI3K) pathway. Loss of PTEN protein expression has been mechanistically linked to tumor progression. Blocking the PI3K pathway might inhibit the growth and proliferation of cells that have deletions in PTEN. Thus characterization of PTEN expression in patient tumor samples may assist prediction of potential response to PI3K inhibitor therapies.

Methods: We developed and validated an immunohistochemical assay on Ventana BenchMark XT to detect PTEN in formalin fixed and paraffin-embedded (FFPE) tissue by utilizing a rabbit monoclonal antibody (clone 138G6) from Cell Signaling Technologies that recognizes the carboxy-terminal domain of PTEN. PTEN immunohistochemical staining was performed on 1577 breast tumor specimens to determine PTEN protein loss. A subset of these cases (n=663) was also assessed for PTEN mutation.

Results: Cellular localization of PTEN expression was observed in both the cytoplasm and the nucleus. Both cellular compartments were scored and used in the staining intensity determination. We observed that PTEN staining is sensitive to variation in tissue handling, fixation and antigen retrieval. PTEN staining was affected significantly by antigen retrieval method. Optimal staining conditions were determined to be 1:60 antibody dilution using Citrate pH 6.0 as antigen retrieval buffer. In a cohort of 1577 hormone receptor positive HER2-negative locally advanced or metastatic breast cancer patients, loss of PTEN protein was observed in 8.6% (135/1577) of patients. There was a positive correlation between PTEN mutation rate (17 out of 66 PTEN IHC positive cases vs 14 out of 587 negative cases) and loss of PTEN protein expression. PTEN loss was observed to correlate with better clinical response to PI3K inhibitor.

Conclusions: Pre-analytical handling of samples is important for PTEN IHC staining. PTEN mutations and insertions/deletions contribute to PTEN protein loss. This study validates a simple method to interrogate PTEN status in clinical specimens and supports the utility of this test in selecting patients who are likely to respond to PI3K inhibitor treatment.

Citation Format: Hua Gong, Emmanuel Pacia, Sharmila Manjeshwar, Beiru Chen, Xun Li, Bashar Dabbas, Jelveh Lameh, Naveen Babbar, Shabnam Tangri. Development and characterization of PTEN IHC assay for testing breast cancer patients specimens [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P1-09-08.

Abstract PD06-03: Development of a highly reproducible clinical test for Ki67 using AQUA technology

Background: The measurement of Ki67 is showing promise in the management of breast cancer patients. It has shown potential for prognosis and as a chemotherapeutic response predictor, and also has a role in multivariate assessment of residual risk. However, the advancement in clinical testing has been arduous due to a lack of clinically validated scoring and/or cutpoints to guide clinicians. Standardization of Ki67 scoring is hindered by inter-laboratory variability in analysis of Ki67 in clinically relevant whole tissue sections, which are complicated not only by the subjective measurement required for IHC but also by heterogeneous expression due to biological variation and inconsistent fixation. In our diagnostic laboratory, we observed a significant amount of variability in the assessment of Ki67 using manual IHC scoring for percentage of positively stained nuclei. To attempt to reduce the variations in Ki67 assessment in clinically relevant specimens, our laboratory investigated the use of image analysis methods. Fluorescence immunohistochemistry (FIHC) coupled with the AQUA® technology has been established as a diagnostic test using image analysis and has been shown to eliminate much of the variation introduced in typical visual assessment.

Materials and methods: FIHC, based on the SP6 antibody clone, was used with AQUA technology to assess Ki67 expression as measured by percentage of positively stained nuclei. Results were compared to manually scored IHC results generated using the commonly used anti-Ki67 MIB1 antibody. IHC interpretation was performed by three pathologists on the same slides.

Results: The Ki67 assay using FIHC and AQUA technology was validated for use in a CLIA environment. Inter-reader variability by IHC was first demonstrated on whole tissues (Average difference in percent positive nuclei was 18% and ranged from 0–50%). This was further assessed on tissue microarrays, which mitigate the heterogeneity component of the tissue and resulted in the same high inter-reader variability (average percent difference was 18%). A significant correlation was observed for Ki67 percent positivity between AQUA and IHC in TMAs (Spearman: r = 0.88, p < 0.0001). FIHC staining and scoring for Ki67 positivity by AQUA resulted in a 3-fold reduced variation compared to IHC and manual reads (6% variation). Using clinically relevant whole tissue sections, the accuracy of Ki67 measurement by AQUA was compared to IHC performed by an outside laboratory and resulted in a significant correlation between the two methods (Spearman: r=0.78, p < 0.0001). The inter-instrument scoring variability by FIHC of a subset of the whole tissue specimens was substantially reduced compared to manual reads (Average difference in percentage of positive nuclei was 2%).

Discussion: The use of image analysis has been shown to be of significant value in expression assessment due to increased objectivity and the potential for broader standardization. We have demonstrated that the use of advanced image analysis when coupled with fluorescence immunohistochemistry can produce a system with greatly improved precision over traditional IHC methods in clinically relevant samples. These results indicate that image analysis of Ki67 is advantageous over manual IHC in clinical diagnostic assessment.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr PD06-03.

Identification and structure-activity relationships of chromene-derived selective estrogen receptor modulators for treatment of postmenopausal symptoms

As part of a program aimed at the development of selective estrogen receptor modulators (SERMs), novel chromene scaffolds, benzopyranobenzoxapanes, were discovered. Many compounds showed binding affinity as low as 1.6-200 nM, displayed antagonist behaviors in the MCF-7 human breast adenocarcinoma cell line as well in Ishikawa cell line with IC(50) values in the range 0.2-360 nM. On the basis of the side chain substitution, various compounds demonstrated strong inhibitory activity in anti-uterotropic assay. Compound 7-(R) and its major metabolites 5-(R) and 6-(R) were evaluated in several in vivo models of estrogen action. Relative to a full estrogen agonist (ethynyl estradiol) and the SERM raloxifene, 7-(R) was found to be a potent SERM that behaved as antagonist in the uterus and exhibited estrogen agonistic activity on bone, plasma lipids, hot flush, and vagina. The overall pharmacokinetic profile and stability were significantly improved compared to those of the phase 2 development compound 9-(R).

Novel chromene-derived selective estrogen receptor modulators useful for alleviating hot flushes and vaginal dryness

A novel SERM (selective estrogen receptor modulators), 1-(R), a chromene-derived bisbenzopyran, was discovered to alleviate hot flushes and effectively increase vaginal fluidity in rats. Moreover, 1-(R) was found to have beneficial effects on plasma cholesterol and bone metabolism while maintaining antiestrogenic activity in the uterus. The biological profile of its enantiomer 1-(S) was also evaluated.

The putative tumor suppressor deleted in malignant brain tumors 1 is an estrogen-regulated gene in rodent and primate endometrial epithelium

Deleted in malignant brain tumors 1 (DMBT1) is a candidate suppressor of malignancies of the brain, lung, gut, and breast. We have been studying gene expression in the uterus in the presence of estrogens and their antagonists. Here, we show that DMBT1 RNA levels are robustly increased by estrogen treatment in the uteri of ovariectomized monkeys and rats. In monkeys, the progestin antagonist mifepristone inhibits estrogen-dependent uterine proliferation. As determined by a microarray experiment and quantitative analysis of RNA levels, mifepristone inhibited estrogenic induction of DMBT1. DMBT1 was not expressed in intact monkeys that were treated with a gonadotropin agonist to suppress steroidogenesis. An in vitro transfection study with human DMBT1 promoter constructs showed that an Alu site approximately 3000 nucleotides upstream of the gene mediates estrogenic regulation. Surprisingly, the estrogen antagonists tamoxifen, raloxifene, and ICI 182,780 also induced gene expression via this Alu site. Rodents represent a more convenient model system for studying uterine biology than monkeys. In rats, uterine DMBT1 RNA levels were dramatically up-regulated by estrogen. Consistent with the transfection study, tamoxifen and raloxifene increased DMBT1 RNA levels in vivo, but ICI 182,780 inhibited an estrogen-induced increase. Immunohistochemical studies showed that DMBT1 is specifically induced in glandular and luminal epithelia of the rat endometrium. Our experiments establish that DMBT1 is an estrogen-responsive gene with a possible role in endometrial proliferation or differentiation, and they have implications for the putative tumor suppressive and mucosal protective functions of DMBT1 in the uterus.

Characterization of human sebaceous cells in vitro

Human sebaceous cells, isolated from adult human skin, were cultured on either bovine type I collagen or mitomycin-C-treated 3T3 fibroblasts. Sebaceous cells, termed "sebocytes", were determined to be epithelial in nature by positive staining with monoclonal antikeratin antibodies BG2 and BG12. However, sebocyte colonies were also negative for keratins found in differentiated cells of keratinocyte colonies, as defined by monoclonal antikeratin antibodies CC2 and CC6. Sebocytes did not produce cornified envelopes in vitro and could only be induced to produce small quantities (less than 5%) of envelopes with a calcium ionophore. Sebocyte growth characteristics in a variety of serum, dexamethasone, and hydrocortisone combinations were significantly different from those of human facial keratinocytes. Sebocytes also displayed a growth curve and plating efficiency that were different from those of keratinocytes. Large lipid droplets within growing sebocytes could be visualized with oil red o staining. Additionally, squalene and wax/cholesterol esters were made by sebocytes in vitro in greater amounts than by facial keratinocytes, as determined by thin-layer chromatography of organic extracts of 3H2O-labeled sebocytes. Sebocytes synthesized greater quantities of lipid, on a per-cell and protein basis, than did keratinocytes.