New treatment options especially of solid tumors including for metastasized prostate cancer (PCa) are urgently needed. Recent treatments of leukemias with chimeric antigen receptors (CARs) underline their impressive therapeutic potential. However CARs currently applied in the clinics cannot be repeatedly turned on and off potentially leading to severe life threatening side effects. To overcome these problems, we recently described a modular CAR technology termed UniCAR: UniCAR T cells are inert but can be turned on by application of one or multiple target modules (TMs). Here we present preclinical data summarizing the retargeting of UniCAR T cells to PCa cells using TMs directed to prostate stem cell- (PSCA) or/and prostate specific membrane antigen (PSMA). In the presence of the respective TM(s), we see a highly efficient target-specific and target-dependent activation of UniCAR T cells, secretion of pro-inflammatory cytokines, and PCa cell lysis both in vitro and experimental mice.

Based on compelling evidence from a vast number of in vitro and in vivostudies, Tregs have become an attractive cell population to treat or even prevent auto- and alloimmunity including Graft-versus-Host disease (GvHD). However, several safety concerns still exist as for example the risk of global immunosuppression using polyclonal Tregs. In fact, experiments in mice showed that adoptive transfer or induction of antigen-specific Tregs is more potent regarding suppression of pathogenic immune responses when compared to polyclonal Treg populations. Unfortunately, the isolation and expansion of naturally occurring antigen-specific Tregs is technically difficult, labour-intensive, and time-consuming. An attractive way to overcome these limitations and to endow polyclonal Treg populations with a desired antigen-specificity is their engraftment with chimeric antigen receptors (CARs). In this context, CAR-modification represents a promising approach to redirect polyclonal Tregs in an antigen-specific manner to suppress ongoing self-destructive immune responses at the site of inflammation. Nevertheless, until now redirection of CAR-engineered T cells is limited to a single target antigen, restricting this approach to an unflexible monospecific therapy. Therefore, we developed a more flexible universal CAR (UCAR) platform that allows redirection of T cells to an in principal unrestricted number of surface antigens. T cells are engrafted with UCARs that bind to a small peptide epitope derived from a human nuclear protein. Cross-linkage to target cells is mediated by independent target modules that provide antigen-specificity and comprise the peptide epitope recognized by the UCAR. In order to target different tissue antigens, the target modules can easily be exchanged. Thereby, once established, the treatment strategy can easily be applied to various auto- and alloimmune diseases. At present, the CD45RA+ population is the Treg subset of choice for a clinical application as these cells have the highest capacity to maintain phenotypic and functional Treg properties upon prolonged ex vivo expansion. Here we show that highly pure, sorted CD4+CD25+CD127lowCD45RA+ Tregs can be genetically manipulated using lentiviral gene transfer, resulting in approximately 70 % of UCAR-expressing Treg cells. The transduction procedure itself did not affect the phenotype of UCAR-engineered Tregs as it was similar to non-transduced wildtype cells. Both Treg populations presevered FOXP3 expression even after prolonged in vitro cultivation (> 95 % FOXP3+). Upon incubation with antigen-positive target cells and a respective target module UCAR-engineered Tregs upregulate the activation markers CD69 and LAP demonstrating that the cells can be restimulated antigen-specifically. Most importantly, UCAR-engrafted Tregs were functionally activated upon antigen encounter, demonstrated by suppression of proliferation and expansion of cocultured autologous T effector cells. Taken together, our results pave the way towards an application of UCAR technology for a site-specific recruitment of CAR-modified Tregs into inflamed tissues aiming at re-establishing immune homeostasis. Due to its high flexibility UCAR-engrafted Tregs can easily and universally be used for treatment of various autoimmune diseases or GvHD just by exchanging the tissue-specific target modules. Disclosures Cartellieri: Cellex Patient Treatment GmbH: Employment. Ehninger:GEMoaB GmbH: Employment, Patents & Royalties. Ehninger:GEMoaB GmbH: Consultancy, Patents & Royalties. Bachmann:GEMoaB GmbH: Consultancy, Patents & Royalties.

Recent treatments of leukemias with T cells expressing chimeric antigen receptors (CARs) underline their impressive therapeutic potential but also their risk of severe side effects including cytokine release storms and tumor lysis syndrome. In case of cross-reactivities, CAR T cells may also attack healthy tissues. To overcome these limitations, we previously established a switchable CAR platform technology termed UniCAR. UniCARs are not directed against typical tumor-associated antigens (TAAs) but instead against a unique peptide epitope: Fusion of this peptide epitope to a recombinant antibody domain results in a target module (TM). TMs can cross-link UniCAR T cells with tumor cells and thereby lead to their destruction. So far, we constructed TMs with a short half-life. The fast turnover of such a TM allows to rapidly interrupt the treatment in case severe side effects occur. After elimination of most of the tumor cells, however, longer lasting TMs which have not to be applied via continous infusion would be more convenient for the patient. Here we describe and characterize a TM for retargeting UniCAR T cells to CD19 positive tumor cells. Moreover, we show that the TM can efficiently be produced in vivo from producer cells housed in a sponge-like biomimetic cryogel and, thereby, serving as an in vivo TM factory for an extended retargeting of UniCAR T cells to CD19 positive leukemic cells.

Combining stem cells with biomaterial scaffolds provides a promising strategy for the development of drug delivery systems. Here we propose an innovative immunotherapeutic organoid by housing human mesenchymal stromal cells (MSCs), gene-modified for the secretion of an anti-CD33-anti-CD3 bispecific antibody (bsAb), in a small biocompatible star-shaped poly(ethylene glycol)-heparin cryogel scaffold as a transplantable and low invasive therapeutic machinery for the treatment of acute myeloid leukemia (AML). The macroporous biohybrid cryogel platform displays effectiveness in supporting proliferation and survival of bsAb-releasing-MSCs overtime in vitro and in vivo, avoiding cell loss and ensuring a constant release of sustained and detectable levels of bsAb capable of triggering T-cell-mediated anti-tumor responses and a rapid regression of CD33+ AML blasts. This therapeutic device results as a promising and safe alternative to the continuous administration of short-lived immunoagents and paves the way for effective bsAb-based therapeutic strategies for future tumor treatments.

Bispecific antibodies (bsAbs) engaging T cells are emerging as a promising immunotherapeutic tool for the treatment of hematologic malignancies. Because their low molecular mass, bsAbs have short half-lives. To achieve clinical responses, they have to be infused into patients continously, for a long period of time. As a valid alternative we examined the use of mesenchymal stromal cells (MSCs) as autonomous cellular machines for the constant production of a recently described, fully humanized anti-CD33-anti-CD3 bsAb, which is capable of redirecting human T cells against CD33-expressing leukemic cells. The immortalized human MSC line SCP-1 was genetically modified into expressing bsAb at sufficient amounts to redirect T cells efficiently against CD33 presenting target cells, both in vitro and in an immunodeficient mouse model. Moreover, T cells of patients suffering from acute myeloid leukemia (AML) in blast crisis eliminated autologous leukemic cells in the presence of the bsAb secreting MSCs over time. The immune response against AML cells could be enhanced further by providing T cells an additional co-stimulus via the CD137-CD137 ligand axis through CD137L expression on MSCs. This study demonstrates that MSCs have the potential to be used as cellular production machines for bsAb-based tumor immunotherapy in the future.

New promising candidates for cancer immunotherapy are bispecific antibodies (bsAbs) which have already shown a convincing anti-tumor effect in first clinical trials. BsAbs are composed of two single-chain fragments variable (scFvs), derived from the variable heavy and light chain of a monoclonal Ab (mAb), which bind to the activating CD3-complex on T cells and a tumor-associated antigen (TAA) on cancer cells. Consequently, the cross-linkage of effector and target cells by bsAbs results in an efficient T cell-mediated cancer cell killing. However, several serious adverse effects were observed in patients after application of Ab constructs. Most critical is an overall activation of T cells which bears the risk of a systemic release of pro-inflammatory cytokines. With respect to the idea that the CD3 binding of an Ab construct is critical for T cell activation, the CD3 domain has to be optimized in order to prevent or even reduce unspecific T cell responses. For this reason, we cloned several anti-CD3 scFvs derived from different anti-CD3 mAb clones and compared their functionality and safety. The anti-CD3 scFv with lowest risk of side effects was introduced in different anti-TAA bsAbs. Functional and safety studies in vitro revealed that development of each bsAb requires extensive and time-consuming optimization steps to gain high efficacy but low risk of side effects. To improve the process of Ab engineering, we recently introduced a novel modular Ab platform that can be rapidly and cost-effectively adapted for redirection of T cells to any TAA (Arndt and Feldmann et al. accepted for publication in Leukemia 2013). In this novel modular system the dual specificity of a conventional bsAb is distributed to two separate Ab modules, (i) the effector module and (ii) the target module. Only the complex of both Ab modules mediates the cross-linkage of effector and target cells resulting in T cell activation and redirected cancer cell lysis similar to conventional T cell engaging bsAbs. The universal effector module is a well optimized bsAb balancing efficacy and safety. It binds CD3 on T cells and the E5B9 tag of the target module. The individual target module comprises an anti-TAA scFv and the peptide epitope E5B9. For treatment of lymphoid or myeloid malignancies a series of different target modules were designed. In vitro and in vivo data clearly underline that the combination of the established effector module with different target modules efficiently activated T cells against hematological malignancies. Both CD4+ and CD8+ T cells from healthy donors were efficiently activated to kill TAA-positive tumor cell lines at low effector to target cell ratios and Ab concentrations in a tumor-specific manner. Most importantly, we could demonstrate that patient-derived T cells were able to kill autologous malignant cells upon Ab-mediated cross-linkage. Another unique feature of the modular system is that multispecific or multifunctional target modules could be easily included to further improve the therapeutic effect. In order to increase anti-tumor specificity and reduce the risk of tumor escape variants, bispecific target modules that recognize different TAAs at the same time, were constructed. Cytotoxicity assays investigating dual targeting of double-positive tumor cells via the modular system demonstrated that target cell lysis can be considerably improved in comparison to single targeting modules. Moreover, proliferation and cytokine release of redirected T cells could be enhanced by supplying costimulatory immunoligands (e.g. 4-1BBL and Ox40L) via target modules. In summary, we developed a multifunctional and highly flexible modular system based on an anti-CD3 domain with lowest risk of side effects. In contrast to conventional anti-TAA-anti-CD3 bsAbs, the development of a novel modular system for different clinical indications is much easier and less time-consuming. Once the effector module containing the critical CD3 domain is optimized, the modular system can be flexibly applied to target any TAA simply by replacing the target module. Application of bispecific target modules or providing costimulatory signals via the modular system might prolong immune responses and further increase anti-tumor specificity and activity. Finally, the novel modular platform represents a valuable and promising tool for cancer immunotherapy. Disclosures: No relevant conflicts of interest to declare.

Despite many years of research and great advances in the field, acute myeloid leukemia (AML) still remains one of the most challenging battle fields in the context of hematologic malignancies treatment. Although AML patients initially respond to conventional chemotherapy, a complete remission is rarely achieved and 5-year survival rates remain low especially in elderly patients. Hence, there is a pressing need for novel effective strategies for AML treatment to prevent relapse and treat minimal residual disease (MRD). The use of recombinant bispecific antibodies (bsAbs) for retargeting effector T lymphocytes towards cancer cells is recently emerging as a promising immunotherapeutic approach for tumor treatment. This class of small molecules is designed to bind simultaneously to a pre-defined tumor-associated antigen (TAA) on tumor cells and the activating CD3 complex on T cells. The cross-linkage of immune effector cell and tumor cell leads to a tumor-specific T cell activation and efficient target cell killing independently of the T cell receptor specificity. However, due to their low molecular mass, bsAbs have a short life span in vivo and consequently have to be continuously administrated to patients over prolonged time spans of several weeks to achieve clinical responses. As an alternative to continuous exogenous infusions of short-lived Abs we examined the use of engineered bone marrow-derived human mesenchymal stem cells (hMSCs) as cellular vehicles for the constant production and secretion of a fully humanized anti-CD33-anti-CD3 bsAb that targets the surface molecule CD33, which is widely overexpressed on AML blasts. Our studies demonstrate that gene-modified hMSCs are effective in releasing the bsAb at sufficient amounts to activate and redirect both human primary CD4+ and CD8+T cells from healthy donors against AML cells expressing varying levels of the CD33 antigen, leading to an efficient T cell-mediated tumor cell killing at low effector to target cell ratios and Ab concentrations. Most importantly, we could demonstrate that patient-derived T cells were able to suppress autologous AML blasts upon Ab-mediated cross-linkage over prolonged period of time without being affected by the presence of the modified hMSCs. Additional improvement of this system was achieved by the artificial expression of T cell co-stimulatory 4-1BB ligand (CD137L) on the hMSCs surface. The additional co-stimulatory signal provided by the engineered hMSCs resulted in an enhanced T cell proliferation, a higher pro-inflammatory cytokine release, and consequently in a more pronounced specific tumor cell killing already at earlier time-points. Taken together, our data could demonstrate that continuous in situ delivery of the anti-CD33-anti-CD3 bsAb by genetically modified hMSCs facilitates efficient activation of T cells for specific and efficient killing of AML blasts over prolonged period of time. Furthermore, as promising perspective of this approach for future in vivo application we are currently investigating on the development of biocompatible synthetic scaffolds as transplantable biomaterial-based production platforms for genetically engineered hMSCs as locally confined vehicle of immunotherapeutics. The implantation of these small engineered devices would ensure that the delivery of the anti-cancer agents can be controlled and stopped after tumor clearance by removing the scaffold at a desired time point. In this way, administration of ex vivo gene-modified hMSCs embedded in appropriate scaffolds would result in a continuous in situ production of recombinant Abs for effective and persistent levels of these therapeutic agents over time with low risk of side effects. Disclosures Cartellieri: Cellex Patient Treatment GmbH, Dresden, Germany: Employment. Ehninger:GEMoaB Monoclonals GmbH, Dresden, Germany: Employment, Patents & Royalties. Ehninger:GEMoaB Monoclonals GmbH, Dresden, Germany: Consultancy, Patents & Royalties. Bachmann:GEMoaB Monoclonals GmbH, Dresden, Germany: Consultancy, Patents & Royalties.

Bispecific T cell engagers (BiTEs) are a growing class of cancer therapy that redirects effector T cells against target-expressing tumors. Mouse models to assess efficacy and cytokine release associated with T cell engagement are needed. JAX humanized NOD scid gamma (Hu-NSG™) mice are engrafted with cord blood-derived CD34+ HSC and have shown robust T cell reconstitution. In this report we evaluated two novel BiTEs, EGFR-CD3 and BCMA-CD3, in Hu-NSG™ mice implanted with MDA-MB-231 human breast cancer cells and NCI-H929 multiple myeloma, respectively. Following the initial dosing, we measured human cytokines and analyzed T cell activation by flow cytometry. Clinical observations and body weight change were also monitored. Treatment with EGFR-CD3 for 2 weeks resulted in complete tumor regression when the dosing was initiated around 100 mm3. 50% of the treated-tumors remained nonpalpable for 10 weeks after dosing ceased. EGFR-CD3 also inhibited tumor growth in mice bearing tumors of ~400 mm3 average in volume. In both cases human IFN-γ, IL-6, and IL-10 levels were elevated from the first dose before returning to the baseline and the frequency of CD69+ cells in T cell subsets increased after the treatment. BCMA-CD3 treatment also induced T cell activation and elevated levels of human IFN-γ, TNF-a, IL-6, and IL-10. NCI-H929 tumor growth inhibition by BCMA-CD3 showed donor to donor variability. We observed that significant growth inhibition in one donor while the other donor did not respond. Together, our Hu-NSG™ platform enables preclinical evaluation of drug efficacy and cytokine release induced by BiTEs and has the potential to be applied to other immunotherapies.