T-cell responses to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) have been described in recovered patients, and may be important for immunity following infection and vaccination as well as for the development of an adoptive immunotherapy for the treatment of immunocompromised individuals. In this report, we demonstrate that SARS-CoV-2–specific T cells can be expanded from convalescent donors and recognize immunodominant viral epitopes in conserved regions of membrane, spike, and nucleocapsid. Following in vitro expansion using a good manufacturing practice-compliant methodology (designed to allow the rapid translation of this novel SARS-CoV-2 T-cell therapy to the clinic), membrane, spike, and nucleocapsid peptides elicited interferon-γ production, in 27 (59%), 12 (26%), and 10 (22%) convalescent donors (respectively), as well as in 2 of 15 unexposed controls. We identified multiple polyfunctional CD4-restricted T-cell epitopes within a highly conserved region of membrane protein, which induced polyfunctional T-cell responses, which may be critical for the development of effective vaccine and T-cell therapies. Hence, our study shows that SARS-CoV-2 directed T-cell immunotherapy targeting structural proteins, most importantly membrane protein, should be feasible for the prevention or early treatment of SARS-CoV-2 infection in immunocompromised patients with blood disorders or after bone marrow transplantation to achieve antiviral control while mitigating uncontrolled inflammation.

Model averaging generally provides better predictions than model selection, but the existing model averaging methods cannot lead to parsimonious models. Parsimony is an especially important property when the number of parameters is large. To achieve a parsimonious model averaging coefficient estimator, we suggest a novel criterion for choosing weights. Asymptotic properties are derived in two practical scenarios: (i) one or more correct models exist in the candidate model set and (ii) all candidate models are misspecified. Under the former scenario, it is proved that our method can put the weight one to the smallest correct model and the resulting model averaging estimators of coefficients have many zeros and thus lead to a parsimonious model. The asymptotic distribution of the estimators is also provided. Under the latter scenario, prediction is mainly focused on and we prove that the proposed procedure is asymptotically optimal in the sense that its squared prediction loss and risk are asymptotically identical to those of the best—but infeasible—model averaging estimator. Numerical analysis shows the promise of the proposed procedure over existing model averaging and selection methods.

The gene regulation network (GRN) is a high-dimensional complex system, which can be represented by various mathematical or statistical models. The ordinary differential equation (ODE) model is one of the popular dynamic GRN models. High-dimensional linear ODE models have been proposed to identify GRNs, but with a limitation of the linear regulation effect assumption. In this article, we propose a sparse additive ODE (SA-ODE) model, coupled with ODE estimation methods and adaptive group least absolute shrinkage and selection operator (LASSO) techniques, to model dynamic GRNs that could flexibly deal with nonlinear regulation effects. The asymptotic properties of the proposed method are established and simulation studies are performed to validate the proposed approach. An application example for identifying the nonlinear dynamic GRN of T-cell activation is used to illustrate the usefulness of the proposed method.

This article studies model averaging for linear mixed-effects models. We establish an unbiased estimator of the squared risk for the model averaging, and use the estimator as a criterion for choosing weights. The resulting model average estimator is proved to be asymptotically optimal under some regularity conditions. Simulation experiments show it is superior or comparable to estimators based on the final models selected by the commonly-used methods and some existing averaging procedures. The proposed procedure is applied to data from an AIDS clinic trial.

Patients with hematologic malignancies relapsing after allogeneic blood or marrow transplantation (BMT) have limited response to conventional salvage therapies, with an expected 1-year overall survival (OS) of <20%. We evaluated the safety and clinical outcomes following administration of a novel T-cell therapeutic targeting 3 tumor-associated antigens (TAA-T) in patients with acute leukemia who relapsed or were at high risk of relapse after allogeneic BMT. Lymphocytes obtained from the BMT donor were manufactured to target TAAs WT1, PRAME, and survivin, which are over-expressed and immunogenic in most hematologic malignancies. Patients received TAA-T infusions at doses of 0.5 to 4 × 107/m2. Twenty-three BMT recipients with relapsed/refractory (n = 11) and/or high-risk (n = 12) acute myeloid leukemia (n = 20) and acute lymphoblastic leukemia (n = 3) were infused posttransplant. No patient developed cytokine-release syndrome or neurotoxicity, and only 1 patient developed grade 3 graft-versus-host disease. Of the patients who relapsed post-BMT and received bridging therapy, the majority (n = 9/11) achieved complete hematologic remission before receiving TAA-T. Relapsed patients exhibited a 1-year OS of 36% and 1-year leukemia-free survival of 27.3% post–TAA-T. The poorest prognosis patients (relapsed <6 months after transplant) exhibited a 1-year OS of 42.8% postrelapse (n = 7). Median survival was not reached for high-risk patients who received preemptive TAA-T posttransplant (n = 12). Although as a phase 1 study, concomitant antileukemic therapy was allowed, TAA-T were safe and well tolerated, and sustained remissions in high-risk and relapsed patients were observed. Moreover, adoptively transferred TAA-T detected by T-cell receptor V-β sequencing persisted up to at least 1 year postinfusion. This trial was registered at clinicaltrials.gov as #NCT02203903.