Citations

Abstract Human immune system (HIS) mice constructed in various ways are widely used for investigations of human immune responses to pathogens, transplants and immunotherapies. In HIS mice that generate T cells de novo from hematopoietic progenitors, T cell-dependent multisystem autoimmune disease occurs, most rapidly when the human T cells develop in the native NOD.Cg- Prkdcscid Il2rgtm1Wjl (NSG) mouse thymus, where negative selection is abnormal. Disease develops very late when human T cells develop in human fetal thymus grafts, where robust negative selection is observed. We demonstrate here that PD-1+CD4+ peripheral (Tph) helper-like and follicular (Tfh) helper-like T cells developing in HIS mice can induce autoimmune disease. Tfh-like cells were more prominent in HIS mice with a mouse thymus, in which the highest levels of IgG were detected in plasma, compared to those with a human thymus. While circulating IgG and IgM antibodies were autoreactive to multiple mouse antigens, in vivo depletion of B cells and antibodies did not delay the development of autoimmune disease. Conversely, adoptive transfer of enriched Tfh- or Tph-like cells induced disease and autoimmunity-associated B cell phenotypes in recipient mice containing autologous human APCs without T cells. T cells from mice with a human thymus expanded and induced disease more rapidly than those originating in a murine thymus, implicating HLA-restricted T cell-APC interactions in this process. Since Tfh, Tph, autoantibodies and LIP have all been implicated in various forms of human autoimmune disease, the observations here provide a platform for the further dissection of human autoimmune disease mechanisms and therapies.

Human immune system (HIS) mice constructed in various ways are widely used for investigations of human immune responses to pathogens, transplants, and immunotherapies. In HIS mice that generate T cells de novo from hematopoietic progenitors, T cell-dependent multisystem autoimmune disease occurs, most rapidly when the human T cells develop in the native NOD.Cg- Prkdc scid Il2rg tm1Wjl (NSG) mouse thymus, where negative selection is abnormal. Disease develops very late when human T cells develop in human fetal thymus grafts, where robust negative selection is observed. We demonstrate here that PD-1 + CD4 + peripheral (Tph) helper-like and follicular (Tfh) helper-like T cells developing in HIS mice can induce autoimmune disease. Tfh-like cells were more prominent in HIS mice with a mouse thymus, in which the highest levels of IgG were detected in plasma, compared to those with a human thymus. While circulating IgG and IgM antibodies were autoreactive to multiple mouse antigens, in vivo depletion of B cells and antibodies did not delay the development of autoimmune disease. Conversely, adoptive transfer of enriched Tfh- or Tph-like cells induced disease and autoimmunity-associated B cell phenotypes in recipient mice containing autologous human APCs without T cells. Tfh/Tph cells from mice with a human thymus expanded and induced disease more rapidly than those originating in a murine thymus, implicating HLA-restricted T cell-APC interactions in this process. Since Tfh, Tph, autoantibodies, and lymphopenia-induced proliferation (LIP) have all been implicated in various forms of human autoimmune disease, the observations here provide a platform for the further dissection of human autoimmune disease mechanisms and therapies.

Objective: We recently discovered that activation of oxidized soluble guanylate cyclase with ataciguat (ATA) slows valve calcification and dysfunction in mice and humans with moderate aortic valve stenosis (AVS). We examined if ATA would attenuate multiple mechanisms contributing to AVS and end organ dysfunction associated with AVS, and hypothesized: 1) ATA attenuates osteogenic signaling in valve tissue from mice with AVS, thereby attenuating calcification and valve dysfunction, 2) AVS increases rates of sudden cardiac death in our model, and 3) treatment with ATA prevents QT prolongation and deleterious electrical remodeling commonly observed in AVS. Methods: Functional, molecular, and histopathological changes in valve tissue along with QT interval duration were measured in 3 independent groups of mice: non-AVS, AVS, and AVS+ATA (WD for 6 months then WD+ATA for 3 months (AVS+ATA). We gave 5 mg/kg isoproterenol (ISO, i.p.) to ldlr -/- /ApoB100-only mice fed normal chow (non-AVS) or western diet (AVS, fed WD for 9 months) to assess risk for malignant arrhythmias. Retrospective analysis of QT interval was also done on data from our recent Phase 1 clinical trial. Results: AVS+ATA mice had less valve calcification and slower progression of valve dysfunction through attenuation of canonical BMP signaling and reductions in Runx2 and SPP1 compared to AVS mice. QT prolongation was evident in AVS mice, and was significantly attenuated by treatment with ATA (p= 0.001) and coarsely associated with severity of valvular or ventricular function. While all non-AVS mice survived for 24 hours after ISO, 40% of mice with AVS reached a moribund equivalent within 4 hours. Retrospective analyses also showed reductions in QT interval duration in patients treated with 200 mg/day ATA for 14 days (p < 0.05). Conclusion: Collectively, this is the first study reporting increased propensity for malignant arrhythmias in our mouse model of AVS, QT prolongation as a likely underlying mechanism, and restoration of cardiac conduction dynamics with targeted restoration of nitric oxide signaling. Future studies focused on the role of sGC in the evolution and prevention of AVS and consequent deleterious electrical remodeling/malignant arrhythmia risk is warranted.

TPS5630 Background: Platinum-resistant ovarian cancer (PROC) portends a poor prognosis and, although treatments are available, response rates are low. Previous studies have demonstrated that DNA topoisomerase I (TOP1) inhibitors exhibit clinical activity in PROC. Irinotecan also has a 15-25% response rate in some studies of platinum-resistant epithelial ovarian cancer but does not have FDA approval for this indication. Additional studies in preclinical models suggest that TOP1 poisons are more active if administered at low doses on prolonged schedules. PLX038 is a long-acting prodrug of the potent TOP1 inhibitor SN-38. In PLX038, SN-38 is covalently conjugated to circulating polyethylene glycol and is slowly released at a constant rate to provide the active agent with a long half-life, a low Cmax and very high exposure – important facets for optimal safety and efficacy of this drug. Further, in preclinical studies the nanoparticle prodrug accumulates and is retained in tumors, where it slowly releases the SN-38. Methods: This single arm, phase 2 clinical trial is open at Mayo Clinic in Rochester, MN. The primary endpoint is overall response rate and secondary endpoints include progression-free survival and assessment of tolerability of PLX038. Correlative goals include i) measurement of TOP1-DNA covalent complexes in tumor biopsies and circulating tumor cells to assess successful target engagement; ii) assessment of homologous recombination status and SLFN11 expression for correlation with tumor response rate; and iii) pharmacokinetics of SN-38 and SN-38G as well as their association with GI toxicity. Assessment of gut microbiota and association with GI toxicity profile are also being investigated. Eligible patients include women with platinum resistant high grade serous ovarian, primary peritoneal, and fallopian tube cancer who have received fewer than two prior chemotherapy regimens in the platinum-resistant setting. All patients must have an ECOG performance status of 0, 1, or 2 and adequate bone marrow, renal, and hepatic function. As of January 31, 2024, 12 of planned 37 patients have enrolled. Clinical trial information: NCT05465941 .

Abstract Neuroendocrine neoplasms are a diverse group of neoplasms that can occur in various areas throughout the body. Well‐differentiated neuroendocrine tumors (NETs) most often arise in the gastrointestinal tract, termed gastroenteropancreatic neuroendocrine tumors (GEP‐NETs). Although GEP‐NETs are still uncommon, their incidence and prevalence have been steadily increasing over the past decades. The primary treatment for GEP‐NETs is surgery, which offers the best chance for a cure. However, because GEP‐NETs are often slow‐growing and do not cause symptoms until they have spread widely, curative surgery is not always an option. Significant advances have been made in systemic and locoregional treatment options in recent years, including peptide‐receptor radionuclide therapy with α and β emitters, somatostatin analogs, chemotherapy, and targeted molecular therapies.

PURPOSE The RECIST guidelines provide a standardized approach for evaluating the response of cancer to treatment, allowing for consistent comparison of treatment efficacy across different therapies and patients. However, collecting such information from electronic health records manually can be extremely labor-intensive and time-consuming because of the complexity and volume of clinical notes. The aim of this study is to apply natural language processing (NLP) techniques to automate this process, minimizing manual data collection efforts, and improving the consistency and reliability of the results. METHODS We proposed a complex, hybrid NLP system that automates the process of extracting, linking, and summarizing anticancer therapy and associated RECIST-like responses from narrative clinical text. The system consists of multiple machine learning–/deep learning–based and rule-based modules for diverse NLP tasks such as named entity recognition, assertion classification, relation extraction, and text normalization, to address different challenges associated with anticancer therapy and response information extraction. We then evaluated the system performances on two independent test sets from different institutions to demonstrate its effectiveness and generalizability. RESULTS The system used domain-specific language models, BioBERT and BioClinicalBERT, for high-performance therapy mentions identification and RECIST responses extraction and categorization. The best-performing model achieved a 0.66 score in linking therapy and RECIST response mentions, with end-to-end performance peaking at 0.74 after relation normalization, indicating substantial efficacy with room for improvement. CONCLUSION We developed, implemented, and tested an information extraction system from clinical notes for cancer treatment and efficacy assessment information. We expect this system will support future cancer research, particularly oncologic studies that focus on efficiently assessing the effectiveness and reliability of cancer therapeutics.

5562 Background: Platinum-resistant or refractory ovarian cancer remains a hard-to-treat disease given the poor prognosis and survival rates. Immune checkpoint inhibitors (ICI; such as pembrolizumab) have not been approved in this indication. Igrelimogene litadenorepvec (Ad5/3-E2F-d24-hTNF-IRES-hIL2; TILT-123), is an oncolytic adenovirus encoding for IL-2 and TNFa, designed to selectively lyse tumor cells while recruiting and activating T-cells, turning immunologically cold tumors hot. The combination of these therapies is appealing as TILT-123 should allow pembrolizumab to be more active, while pembrolizumab should delay emergence of T-cell exhaustion. Methods: PROTA (NCT05271318) is a multicenter open-label phase 1 clinical trial initiated to investigate the safety of TILT-123 in combination with pembrolizumab, in participants bearing ovarian cancer resistant or refractory to platinum-based therapy. In a 3+3 dose-escalation design, TILT-123 was administered intravenously and intratumorally/intraperitoneally to participants with an injectable tumor or carcinomatosis. Pembrolizumab (200mg/Q3wk) was not altered. Eligibility criteria included adequate hepatic and renal functions, and WHO/ECOG performance status of 0-1. Primary endpoint was safety by day 92 evaluated based on adverse events, laboratory values, vital signs, and electrocardiogram. Secondary endpoints included anti-tumor response (RECIST1.1/iRECIST), CA125 levels, pharmacokinetics, transduction of tumors, immune response towards the virus and tumors. Data cutoff was 22Jan2024. Results: 15 patients were enrolled with a median age of 66 (36-78) years, and 14 were evaluable for treatment response with RECIST 1.1/iRECIST. 8 patients were platinum-resistant, and 7 patients were platinum-refractory. Majority of patients had WHO/ECOG performance status 1. Prevalent cancer histology was epithelial ovarian cancer, with some fallopian tube cancer and primary peritoneal cancer. Dose-limiting toxicities were not observed. Frequent grade 1-4 treatment-related AEs (>30% of patients) were anemia, nausea, lymphocyte count decreased, fever and diarrhea. Stable disease (SD) by RECIST1.1 was achieved in 64.3% of evaluable patients and CA125 decreased below baseline in 50% of the patients at least in one timepoint. Median PFS and OS in all patients were 105 and 280 days, respectively. In patients with SD on day 92 by RECIST1.1, median OS was 293 days and median PFS was 174 days. Tumor biopsies indicated the presence of virus and induction of T-cells in injected and non-injected lesions. Conclusions: Combination of TILT-123 and pembrolizumab is safe and appears to induce disease control in a difficult-to-treat patient population. Investigation of TILT-123, pembrolizumab and pegylated liposomal doxorubicin is ongoing in similar patients. Clinical trial information: NCT05271318 .

11145 Background: T-DXd is increasingly used in mBC. A T-DXd toxicity of special interest is interstitial lung disease/pneumonitis (ILD), occurring in 10-15% of pts in the DESTINY-Breast trials, and with potential lower incidence/severity in earlier line settings [Krop et al, ASCO 2023]. However, in routine practice, T-DXd may be used in pts with more heavily pre-treated mBC and previously exposed to other antibody-drug conjugates (ADC) [Premji et al SABCS 2023]. We evaluated the incidence/severity of T-DXd related ILD in mBC at Mayo Clinic. Methods: We retrospectively identified pts with mBC who received ≥1 dose of T-DXd at Mayo Clinic, Rochester, MN, between July 2022-December 2023. Demographic, mBC, and pulmonary clinical variables were abstracted from the clinical records. Data was summarized using descriptive statistics. Diagnosis of ILD was determined by treating clinicians, and severity to CTCAE V5 based on clinical documentation. Results: 77 pts with mBC received T-DXd during the study period. All were female, the majority (70, 91%) were Caucasian, and median age was 58. 53 (69%) had HER2-low mBC (37 HR+, 16 HR-) and 24 (31%) had HER2+ mBC (17 HR+, 7 HR-). 31 (40%) were active or former smokers. 11 (14%) had pulmonary comorbidities, including 6 (7.8%) with asthma, 2 (2.6%) with prior pneumonitis, and 3 (3.9%) with other pulmonary comorbidities. They had previously received a median of 1 line of endocrine therapy, and 4 lines of chemotherapy, with 20 (26%) having received another ADC. 17 (22%) developed any grade ILD [G1: 3 (4%), G2: 7 (9%), G3: 2 (3%), G4: 1 (1%), G5: 4 (5%)]. Among pts with ILD of any grade, 13 (76.5%) presented with at least 1 symptom (cough/SOB). The median number of T-DXd cycles prior to onset of any grade ILD was 6 (range 2-13), or 18 (range 6-39) weeks. 5 (29%) pts with any grade ILD received a prior ADC. Of pts with ≥G3 ILD (n=7), all were Caucasian, 2 (30%) were prior smokers, none had prior lung disease, and had received a median of 5 prior lines of chemo, with 4 (57%) having received a prior ADC. All grade ≥2 ILDs were treated with systemic steroids, and 1 also received IVIG. CT chest was used for diagnosis of all ILD. 7 pts (41%) required hospitalization, all of whom had bronchoscopy, with 5 (29%) requiring intubation. All pts with G1 ILD (3, 18%) restarted T-DXd without ILD recurrence. Conclusions: In this case series,22% of ps treated with T-DXd experienced ILD (5% G5), higher than the rate observed in the pivotal DESTINY-Breast trials, possibly due to real-world use of T-DXd in more heavily pre-treated pts and in pts with pulmonary comorbidities.

513 Background: Adjuvant endocrine therapy (AET) improves overall survival (OS) in patients (pts) with hormone receptor (HR) positive early-stage breast cancer (BC). For pts with estrogen receptor (ER)-low BC (defined as ER 1-10%), there is controversy regarding AET benefit, given tumor heterogeneity (>40% basal; (1)), and similar chemotherapy response and prognosis to ER-negative (neg) BC. Current guidelines categorize ER-low BC as a separate entity and suggest equipoise regarding AET benefit. We previously determined that AET omission in chemotherapy-treated pts with ER+ BC resulted in a nearly 2-fold increase in death (2). However, the impact of AET omission in patients with ER-low BC is unknown. Methods: Beginning in 2018, the National Cancer Database (NCDB) began to abstract ER as a continuous variable, allowing classification of ER+ tumors as 1-10% or >10%. Therefore, we queried the NCDB 2018-2020 for patients with stage I-III ER+BC treated with neoadjuvant or adjuvant chemotherapy, as these pts have the highest risk of recurrence. ER was assessed as ER-neg (0%), low (1-10%) and ER >10%. OS was analyzed with AET as a time-dependent covariate using Cox proportional hazards regression. Results: 357,085 pts with stage I-III ER+BC were identified, with 10,396 (3%) classified as ER-low. Among pts with ER-low disease, 69% had PR-neg BC and 67% had HER2-neg BC. The majority (77%) were treated with chemotherapy (32% neoadjuvant, 44% adjuvant, 1% unknown timing) and formed our analysis cohort. Among these 7,956 pts, AET was omitted in 3,233 (41%). AET omission was more common for pts with tumors that were PR-neg vs positive (46% vs 30%, p<0.001), HER2-neg vs positive (44% vs 37%, p<0.001), and Ki67≥20% (44% vs 35%, p<0.001). With median follow-up of 3 years, 648 deaths were observed. OS was 94.3% (95% CI: 93.8-94.9%) at 2 years and 87.6% (95% CI: 86.5-88.7%) at 4 years. In unadjusted analysis, omission of AET was associated with worse OS (HR 1.40, 95% CI: 1.19-1.65, p<0.001), with similar effects regardless of PR, HER2, or Ki67 (each interaction test p>0.3). When controlling for age, comorbidity score, year of diagnosis, PR, HER2, and pathologic stage, the adjusted effect of AET omission on OS was HR 1.22 (95% CI: 1.00-1.48, p=0.05). Given assessment of OS was performed during pandemic years, with 1/3 of deaths occurring during the first year of follow-up, we performed a sensitivity analysis to analyze pts who survived to one year after definitive surgery. In this analysis, the adjusted effect of AET omission on OS was 1.24 (95% CI: 1.02-1.51, p=0.03). Conclusions: Omission of AET in pts treated with chemotherapy for ER-low, early-stage BC is associated with significantly worse OS. These data strongly suggest that pts with ER low BC should be counseled regarding the benefit of AET and practice guidelines should recommend AET in this setting. 1. Benefield JNCI 2020. 2. Choong SABCS 2022.