Knutson, Keith L.

Abstract Ligation of programmed cell death-1 (PD-1) in the tumor microenvironment is known to inhibit effective adaptive antitumor immunity. Blockade of PD-1 in humans has resulted in impressive, durable regression responses in select tumor types. However, durable responses have been elusive in ovarian cancer patients. PD-1 was recently shown to be expressed on and thereby impair the functions of tumor-infiltrating murine and human myeloid dendritic cells (TIDC) in ovarian cancer. In the present work, we characterize the regulation of PD-1 expression and the effects of PD-1 blockade on TIDC. Treatment of TIDC and bone marrow–derived dendritic cells (DC) with IL10 led to increased PD-1 expression. Both groups of DCs also responded to PD-1 blockade by increasing production of IL10. Similarly, treatment of ovarian tumor–bearing mice with PD-1 blocking antibody resulted in an increase in IL10 levels in both serum and ascites. While PD-1 blockade or IL10 neutralization as monotherapies were inefficient, combination of these two led to improved survival and delayed tumor growth; this was accompanied by augmented antitumor T- and B-cell responses and decreased infiltration of immunosuppressive MDSC. Taken together, our findings implicate compensatory release of IL10 as one of the adaptive resistance mechanisms that undermine the efficacy of anti–PD-1 (or anti–PD-L1) monotherapies and prompt further studies aimed at identifying such resistance mechanisms. Cancer Res; 77(23); 6667–78. ©2017 AACR.

Abstract Background: Trastuzumab (H), a monoclonal antibody against HER2, has revolutionized the treatment for HER2-positive breast cancer. Besides inhibiting downstream signaling of HER2, several studies showed that H also exerts its anti-tumor activity via immune-related mechanisms. While H is quite effective in preventing recurrence, significant numbers of patients still develop recurrence despite adjuvant H-based chemotherapy. In this study, we analyzed immune-related genes that were associated with poor outcome in N9831. Methods: NanoStringTM technology was used to quantify mRNA of immune-related genes in baseline samples from 1,280 patients in N9831. In N9831, patients in arm A were treated with chemotherapy alone (AC-T), arm B received chemotherapy followed by sequential H (AC-T-H), and arm C received H concurrently with chemotherapy (AC-TH). Cox proportional hazard ratio (HR) was used to determine the association of each immune-related gene with recurrence free survival (RFS). Different immune subset signatures, including CD45, CD8, cytotoxic-cells, and T-cells were analyzed using algorithms developed by NanoString. Results: With the median follow up of 10.6 years, we identified a total 77 genes that were associated with improved outcome in arm C. Among these 77 genes, there were 20 tumor necrosis factor (TNF)-related genes. Of those, only OX40 (TNFRSF4) and its ligand TNFSF4 have interaction p < 0.10. Interestingly, we found uneven distribution of OX40 expression in the N9831 specimens. Approximately, 9.5% of HER2-positive breast cancer expressed OX40 at distinctly low level. Low expression of OX40 was significantly associated with HR positivity (OX40 low 61% vs. OX40 high 51%, p 0.003) and larger tumor size but not patients’ age, tumor grade, and lymph node status. Low expression of OX40 was significantly associated with low mTIL-CD45, CD8, cytotoxic-cells, and T-cells immune signature scores. Tumors with low OX40 expression had significantly lower levels of CTLA4 (p 9.11e-71) and PD-L1 expressions (p 2.49e-92). Low expression of OX40 is associated with poorer outcome among patients treated with H in both sequential (HR 0.88, 95% CI 0.79-0.98, p 0.022) and concurrent arms (HR 0.86, 95%CI 0.76-0.98, p 0.027) but not in chemotherapy only arm (HR 0.9, 95%CI 0.90-1.09, p 0.8). Similar findings were observed with its ligand TNFSF4 in arm A (HR 1.02, 95%CI 0.92-1.15, p 0.68), arm B (HR 0.87, 95%CI 0.76-0.99, p 0.04), and arm C (HR 0.81, 95%CI 0.68-0.95, p 0.01). Conclusion: Our study suggests that pre-existing expression of OX40 and its ligand TNFSF4 are prognostic and may also be predictive of adjuvant H benefit. Patients with distinctly low level of OX40 had poor outcome despite adjuvant H-based adjuvant chemotherapy. Our study provides a rationale to further evaluate the strategy to increase immune activation to improve outcome in this group of patients. Citation Format: Saranya Chumsri, Daniel J. Serie, Afshin Mashadi-Hossein, Sarah Warren, Alvaro Moreno-Aspitia, Geraldo Colon-Otero, Keith L. Knutson, Edith A. Perez, E. Aubrey Thompson. Correlation of immune co-stimulatory molecule OX40 and outcome in trastuzumab treated HER2-positive breast cancer patients in the NCCTG-N9831 [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 1682. doi:10.1158/1538-7445.AM2017-1682

Abstract Purpose: Little is known about the role of the immune system in the earliest stages of breast carcinogenesis. We studied quantitative differences in immune cell types between breast tissues from normal donors and those from women with benign breast disease (BBD). Experimental Design: A breast tissue matched case–control study was created from donors to the Susan G. Komen for the Cure Tissue Bank (KTB) and from women diagnosed with BBD at Mayo Clinic (Rochester, MN) who either subsequently developed cancer (BBD cases) or remained cancer-free (BBD controls). Serial tissue sections underwent immunostaining and digital quantification of cell number per mm2 for CD4+ T cells, CD8+ T cells, CD20+ B cells, and CD68+ macrophages and quantification of positive pixel measure for CD11c (dendritic cells). Results: In 94 age-matched triplets, BBD lobules showed greater densities of CD8+ T cells, CD11c+ dendritic cells, CD20+ B cells, and CD68+ macrophages compared with KTB normals. Relative to BBD controls, BBD cases had lower CD20+ cell density (P = 0.04). Nearly 42% of BBD cases had no CD20+ B cells in evaluated lobules compared with 28% of BBD controls (P = 0.02). The absence of CD20+ cells versus the presence in all lobules showed an adjusted OR of 5.7 (95% confidence interval, 1.4–23.1) for subsequent breast cancer risk. Conclusions: Elevated infiltration of both innate and adaptive immune effectors in BBD tissues suggests an immunogenic microenvironment. The reduced B-cell infiltration in women with later breast cancer suggests a role for B cells in preventing disease progression and as a possible biomarker for breast cancer risk. Clin Cancer Res; 23(14); 3945–52. ©2017 AACR.

Abstract PD-1 and PD-L1 form a major inhibitory axis that acts to suppress tumor-rejecting effector responses. In addition to its expression and inhibitory functions in T and B cells, PD-1 has also been shown to be expressed on and hence impact the functions of cells of innate arm, such as dendritic cells (DCs), primarily at the tumor site but not in the periphery. Our aim in this study was to identify tumor-associated factor(s) responsible for PD-1 regulation on DCs and investigate how DCs in the tumor microenvironment respond to anti-PD-1-based immunotherapy. Both bone marrow-derived dendritic cells (BMDCs) as well as murine ovarian tumor ascites-derived DCs were used in the studies. Using in vitro cultures, we show that the cytokine IL-10; which is expressed in abundance in many malignancies including ovarian cancer, is potent regulator of PD-1 expression on DCs. Based on inhibition and siRNA knockdown studies; we show that IL-10 mediated PD-1 expression depends on STAT3 activation. Treatment of DCs with IL-10 also led to an increase in expression of PD-L1 on surface of DCs as well as an increase in release of soluble PD-L1. Antibody-mediated blockade of PD-1 on DCs led to greatly increased production of IL-10 (> 4 fold) in cultures in vitro. Furthermore, treatment of tumor bearing mice with anti-PD-1 antibody led to a significant increase of IL-10 in ascites (> 6 fold) and periphery (> 4 fold). This compensatory release of IL-10 correlated with a further increase in PD-1 expression on DCs both in vitro and in vivo. In in vivo studies, while the blockade of PD-1 or IL-10 neutralization as monotherapy were ineffective, blockade of PD-1 and IL-10 neutralization as combination therapy augmented the anti-tumor response in ovarian tumor bearing mice; leading to a decrease in tumor growth and a significant increase in survival. These results show that the IL-10 and PD-1 pathways intersect with DCs in the tumor microenvironment. PD-1 blockade on DCs leads to compensatory release of IL-10 resulting in the maintenance of immunosuppression, and a combination regimen of PD-1 blockade and IL-10 neutralization has therapeutic benefit in ovarian tumors. These results serve as catalyst to explore the feasibility/efficacy of exploiting these targets in clinical settings for treatment of human ovarian cancer. Citation Format: Purushottam Lamichhane, Lavakumar Karyampudi, Barath Shreeder, James Krempski, Deborah Bahr, Matthew Block, Keith Knutson. Blockade of PD-1 signaling in tumor-associated dendritic cells results in compensatory IL-10 release maintaining immune suppression in ovarian cancer microenvironments. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research: Exploiting Vulnerabilities; Oct 17-20, 2015; Orlando, FL. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(2 Suppl):Abstract nr B66.

Abstract The addition of trastuzumab to chemotherapy extends survival among patients with HER2+ breast cancer. Prior work showed that trastuzumab and chemotherapy augments HER2 extracellular domain (ECD)-specific antibodies. The current study investigated whether combination therapy induced immune responses beyond HER2-ECD and, importantly, whether those immune responses were associated with survival. Pretreatment and posttreatment sera were obtained from 48 women with metastatic HER2+ breast cancer on NCCTG (now Alliance for Clinical Trials in Oncology) studies, N0337 and N983252. IgG to HER2 intracellular domain (ICD), HER2-ECD, p53, IGFBP2, CEA, and tetanus toxoid were examined. Sera from 25 age-matched controls and 26 surgically resected HER2+ patients were also examined. Prior to therapy, some patients with metastatic disease had elevated antibodies to IGFBP2, p53, HER2-ICD, HER2-ECD, and CEA, but not to tetanus toxin, relative to controls and surgically resected patients. Treatment augmented antibody responses to HER2-ICD in 69% of metastatic patients, which was highly associated with improved progression-free survival (PFS; HR = 0.5, P = 0.0042) and overall survival (OS; HR = 0.7, P = 0.038). Augmented antibody responses to HER2-ICD also correlated (P = 0.03) with increased antibody responses to CEA, IGFBP2, and p53, indicating that treatment induces epitope spreading. Paradoxically, patients who already had high preexisting immunity to HER2-ICD did not respond to therapy with increased antibodies to HER2-ICD and demonstrated poorer PFS (HR = 1.6, P < 0.0001) and OS (HR = 1.4, P = 0.0006). Overall, the findings further demonstrate the importance of the adaptive immune system in the efficacy of trastuzumab-containing regimens. Cancer Res; 76(13); 3702–10. ©2016 AACR.

Abstract The PD-1:PD-L1 immune signaling axis mediates suppression of T-cell–dependent tumor immunity. PD-1 expression was recently found to be upregulated on tumor-infiltrating murine (CD11c+CD11b+CD8−CD209a+) and human (CD1c+CD19−) myeloid dendritic cells (TIDC), an innate immune cell type also implicated in immune escape. However, there is little knowledge concerning how PD-1 regulates innate immune cells. In this study, we examined the role of PD-1 in TIDCs derived from mice bearing ovarian tumors. Similar to lymphocytes, TIDC expression of PD-1 was associated with expression of the adapter protein SHP-2, which signals to NF-κB; however, in contrast to its role in lymphocytes, we found that expression of PD-1 in TIDC tonically paralyzed NF-κB activation. Further mechanistic investigations showed that PD-1 blocked NF-κB–dependent cytokine release in a SHP-2–dependent manner. Conversely, inhibition of NF-κB–mediated antigen presentation by PD-1 occurred independently of SHP-2. Collectively, our findings revealed that PD-1 acts in a distinct manner in innate immune cells compared with adaptive immune cells, prompting further investigations of the signaling pathways controlled by this central mediator of immune escape in cancer. Cancer Res; 76(2); 239–50. ©2015 AACR.

Abstract Introduction: CD56+ lymphocytes (a defining marker of natural killer cells) have a predominantly cytotoxic phenotype and a hypothetical role in breast tumor immunosurveillance. Here we investigate whether CD56+ staining density varies in nonmalignant breast lobules according to confirmed breast cancer risk factors- age, and histologic features of fibrocystic change and lobular involution. Methods: Archived breast tissue samples were obtained from 94 women with benign breast disease (BBD). The sample set was selected as a nested case-control study (47 cases, 47 controls) from within a large BBD cohort; cases were those who developed breast cancer subsequent to their benign breast biopsy; controls were matched to cases on age at biopsy, year of biopsy, and length of breast cancer free follow-up at least as long as the time-to-cancer in the matched case. Up to 10 representative lobules in each sample were characterized by H&E for fibrocystic changes and the degree of epithelial proliferation (normal, nonproliferative, proliferative, atypia) and for degree of lobular involution (none, partial, complete). Consecutive sections were immunostained for CD56 and hematoxylin. Using digital image analysis, CD56 staining was quantified on a per lobule basis by pixel ratio (PR) (positive stain:total lobule). Statistical analysis was performed using linear mixed effects regression on the rank transformed density to account for correlations among multiple lobules from the same patient in assessing the association of density with lobule characteristics and using conditional logistic regression for association with case-control status. Results: Among 94 women (median age 52, range 35-73), 876 lobules were evaluated: 431 in BBD cases and 445 in BBD controls. Overall, 391 lobules (45%) were normal and the remaining 485 (55%) demonstrated fibrocystic changes. Among fibrocystic lobules, 247 (51%) had nonproliferative changes, 218 (45%) had epithelial proliferation, and 20 (4%) had atypical hyperplasia. Among normal lobules, 61 (16%) had no involution, 139 (36%) had partial involution, and 191 (49%) had complete involution. The vast majority of lobules (861/876, 98%) demonstrated presence of CD56+ cells, while only 15 of 876 lobules in 11 patients had zero CD56+ cells [7 cases (15%) versus 4 controls (8.5%) OR 1.75, p = 0.37]. The median CD56+ PR was 0.55% overall. Fibrocystic lobules as a group showed significantly lower median CD56+ PR compared to normal lobules (0.37% vs 0.85%, p < 0.0001). Among subcategories of fibrocystic lobules, the median CD56+ PR was lower in proliferative lobules (0.29%) compared to non-proliferative (0.42%), but this difference was not statistically significant (p = 0.55). Older women (age > 55) showed the highest CD56+ PR (median of 0.92%), as compared to 0.38% in women 45-55 (p=0.02), and 0.52% in women age <45 (p = 0.26). Among normal lobules, median CD56+ PR was similar between lobules with no involution (0.55%) and partial involution (0.52%) but was considerably higher among lobules with complete involution (1.3%, p = 0.09 and p = 0.05, respectively). In a multivariable model, CD56+ PR differences between fibrocystic and normal lobules remained significant (p = 0.0002) and also remained significant between age groups of > 55 and 45-55 (p = 0.02), while involution was non-significant. Median CD56+ PR was lower in cases (0.43%) compared to controls (0.72%), but this was not statistically significant (p=0.19) in our sample of 47 case-control pairs. Conclusions: Breast lobules with fibrocystic changes show less CD56+ cells, raising a question of possible immune suppression induced by early epithelial abnormalities. These findings encourage further investigation to improve understanding of the immune microenvironment in premalignant breast tissues. Citation Format: Rushin D. Brahmbhatt, Daniel Kerekes, Tanya L. Hoskin, Alvaro Pena, Daniel W. Visscher, Melody L. Stallings-Mann, Derek C. Radisky, Linda M. Murphy, Vernon Shane Pankratz, Keith L. Knutson, Marlene Frost, Amy C. Degnim. CD56+ immune cell infiltration is decreased in benign breast lobules with fibrocystic changes. [abstract]. In: Proceedings of the Thirteenth Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2014 Sep 27-Oct 1; New Orleans, LA. Philadelphia (PA): AACR; Can Prev Res 2015;8(10 Suppl): Abstract nr B34.

Abstract Purpose: It is well known that the immune system impacts the clinical course of ovarian cancer suggesting that immune-based approaches may effective in treating the disease. This has led to several clinical trials of novel treatments such as antibody, vaccine, and adoptive T cell therapy. PD-1:PD-L1 axis is a major immune regulatory pathway that blunts immune effectors in the tumor microenvironment. It is well known that high surface expression of PD-1 on tumor infiltrating T cells is a sign of their exhaustion. However, the complete mechanism behind the PD-1 regulation of T cell exhaustion is yet to be explored. Recently we identified that PD-1 is also expressed on ovarian tumor-infiltrating myeloid dendritic cells (DCs) that exhibit an immunosuppressive phenotype. Our goal in this study was to understand the mechanism by which PD-1 mediates the paralysis of ovarian cancer infiltrating DCs. Experimental procedures: PD-1+ DCs obtained from ID8 mouse model of ovarian cancer were used in this study. Data from this preclinical model was confirmed by using myeloid DCs obtained from ovarian cancer patient’s samples. Using standard flow cytometry, immunoassays (ELISA, multiplexed cytokine assay), PCR array, western blot and confocal microscopy, PD-1 mediated paralysis of ovarian cancer DCs was determined. Results: Our data shows that blockade of surface PD-1 on ovarian cancer infiltrating DCs results in their activation by, enhancing their surface co-stimulatory molecules expression, increasing their motility, enhancing their antigen presentation capacity, and increasing their production of immunostimulatory cytokines. Also, our results suggest that PD-1 expressed on tumor DCs mediates the suppression of NFkB which occurs through SHP-2 and IKK dependent mechanisms. Conclusion: Our study reveals that blockade of PD-1 on ovarian cancer infiltrating DCs results in the reversal of their paralysis to such an extent that it overrides their B7-H1 mediated suppression of T cells and this occurs through the reversal of PD-1 mediated tonic suppression of NFkB. Citation Format: Lavakumar Karyampudi James Krempski, Purushottam Lamichhane, Kimberly R. Kalli, Marshall D. Behrens, Doris M. Vargas, Lynn C Hartmann, Karen E Hedin, Ellen L. Goode, Keith L. Knutson. PD-1 mediated paralysis of ovarian cancer infiltrating dendritic cells [abstract]. In: Proceedings of the 10th Biennial Ovarian Cancer Research Symposium; Sep 8-9, 2014; Seattle, WA. Philadelphia (PA): AACR; Clin Cancer Res 2015;21(16 Suppl):Abstract nr POSTER-THER-1418.