Multicenter study of banked third-party virus-specific T cells to treat severe viral infections after hematopoietic stem cell transplantation

Abstract

Virus-specific T cell (VST) lines could provide useful antiviral prophylaxis and treatment of immune-deficient patients if it were possible to avoid the necessity of generating a separate line for each patient, often on an emergency basis. We prepared a bank of 32 virus-specific lines from individuals with common HLA polymorphisms who were immune to Epstein-Barr virus (EBV), cytomegalovirus, or adenovirus. A total of 18 lines were administered to 50 patients with severe, refractory illness because of infection with one of these viruses after hematopoietic stem cell transplant. The cumulative rates of complete or partial responses at 6 weeks postinfusion were 74.0% (95% CI, 58.5%-89.5%) for the entire group (n = 50), 73.9% (95% CI, 51.2% -96.6%) for cytomegalovirus (n = 23), 77.8% for adenovirus (n = 18), and 66.7% (95% CI, 36.9%-96.5%) for EBV (n = 9). Only 4 responders had a recurrence or progression. There were no immediate infusion-related adverse events, and de novo graft-versus-host disease developed in only 2 patients. Despite the disparity between the lines and their recipients, the mean frequency of VSTs increased significantly postinfusion, coincident with striking decreases in viral DNA and resolution of clinical symptoms. The use of banked third-party VSTs is a feasible and safe approach to rapidly treat severe or intractable viral infections after stem cell transplantation. This study is registered at www.clinicaltrials.gov as NCT00711035.

Figures

Figure 1

Cumulative response rates at 6 weeks postinfusion of VSTs. (A) Results based on the first CR/PR in all 50 patients, regardless of the type of infection. (B) Results by specific viral infection: CMV, AdV, and EBV. The curves were constructed with use of the cumulative incidence function. Further details are given in “Statistical analysis.”

Figure 2

Outcomes after treatment with banked VSTs: CMV-infected patients. (A) Depiction of viral load measurements before and during 6 weeks postinfusion of VSTs in 19 evaluable patients (responders and nonresponders) treated for persistent CMV infection. Results are presented as copies per milliliter (CMV detected in blood) or as copies per cell in patients with CMV antigenemia. Because viral loads were measured by different assays at different institutions, interpatient comparisons should not be made. (B) An example of a patient with CMV colitis with evidence of viral inclusions on H&E staining that contained CMV by immunostaining (left panel). After VST infusion, there was a complete resolution of CMV colitis and no evidence of viral inclusions by H&E staining (right panel). (C) Frequency of CMVpp65–directed T cells in peripheral blood before and at weeks 1, 2, 4, and 12 postinfusion, as measured with an IFN-γ ELIspot assay after overnight stimulation of PBMCs with CMVpp65 pepmix. Results are expressed as mean (±SD) spot-forming cells (SFCs) per 2 × 105 input cells.

Figure 2

Outcomes after treatment with banked VSTs: CMV-infected patients. (A) Depiction of viral load measurements before and during 6 weeks postinfusion of VSTs in 19 evaluable patients (responders and nonresponders) treated for persistent CMV infection. Results are presented as copies per milliliter (CMV detected in blood) or as copies per cell in patients with CMV antigenemia. Because viral loads were measured by different assays at different institutions, interpatient comparisons should not be made. (B) An example of a patient with CMV colitis with evidence of viral inclusions on H&E staining that contained CMV by immunostaining (left panel). After VST infusion, there was a complete resolution of CMV colitis and no evidence of viral inclusions by H&E staining (right panel). (C) Frequency of CMVpp65–directed T cells in peripheral blood before and at weeks 1, 2, 4, and 12 postinfusion, as measured with an IFN-γ ELIspot assay after overnight stimulation of PBMCs with CMVpp65 pepmix. Results are expressed as mean (±SD) spot-forming cells (SFCs) per 2 × 105 input cells.

Figure 3

Outcomes after treatment with banked VSTs: AdV-infected patients. (A) Serial viral load measurements before and during 6 weeks post-VST infusion in 17 evaluable patients (responders and nonresponders) treated for persistent AdV. Results are presented as copies per milliliter, where Adv was detected either in blood or in the stool. Because viral loads were measured by different assays at different institutions, interpatient comparisons should not be made. In (B) and (C), elevated AdV loads are detected in both blood (B) and stool (C) in a patient (#59) whose infection resolved after VST infusion. Results are expressed as copies per milliliter. (D) Frequency of AdV-directed T cells in peripheral blood before and at weeks 1, 2, 4, and 6 postinfusion, as measured with an IFN-γ ELIspot assay after overnight stimulation of PBMCs with Adv pepmixes (Hexon and Penton). Results are expressed as mean (±SD) SFCs per 2 × 105 input cells.

Figure 4

Outcomes after treatment with banked VSTs: EBV-infected patients. (A) Serial viral load measurements before and during 6 weeks post-VST infusion in patients (responders and nonresponders) with EBV-LPD. Results are presented as copies per milliliter. (B, left) Positron emission tomography scan of a patient with extensive and progressive LPD despite sirolimus therapy, with a second scan (right) showing complete resolution of all lesions after VST infusion. (C) Frequency of EBV-directed T cells in peripheral blood before and at weeks 1, 2, and 4 postinfusion, as measured with an IFN-γ ELIspot assay after overnight stimulation of PBMCs with EBV lymphoblastoid cell lines. Results are expressed as mean (±SD) SFC per 2 × 105 input cells.