Adoptive T-cell Therapy with CD45RA-depleted Donor in the Treatment of Cytomegalovirus Disease in Immunocompromised Non-transplant Patients

Introduction

Cytomegalovirus (CMV) is a major opportunistic pathogen frequently reactivated from latency during prolonged immunosuppression [1]. Currently, ganciclovir or valganciclovir are considered the preferred therapeutic options. When toxicity or resistance arise, foscarnet, cidofovir, leflunomide and artesunate and more recently new antivirals such as maribavir, letermovir or brincidofovir could be used [2].

Cell-mediated immunity essentially controls both CMV infection and disease [3,4]. Infusion of CMV-specific T-cells in CMV disease refractory to antivirals has been used after haematopoietic stem cell transplantation [5]. Clinical experience in solid organ transplant recipients is limited to three patients, with successful outcome in two cases [6–8]. Adoptive immunotherapy in non-transplant patients has not been described.

There are several strategies to obtain CMV-specific T-cells. These cells can be generated in vitro or isolated by tetramers [9] or using immunomagnetic selection of T-cells that secrete interferon (IFN)-γ after stimulation with viral proteins or peptides [10]. This last technique offers simultaneous selection of both CD4⁺ and CD8⁺ cells and makes it possible to obtain specific and functional cells. For isolation, it is necessary a donor with specific virus T-cells against the peptides is available to make the selection. An alternative to the selection of T virus specific cells is the use of depleted T-cells of CD45RA. CD45RA is a membrane molecule present on some B-cells and natural killer (NK) cells, as well as on plasmocytic dendritic cells and some CD34⁺ haematopoietic stem cells but mainly, CD45RA is expressed on naive T-cells, whereas memory T-cells are CD45RA negative (CD45RA-). CD45RA depletion results in a cellular product passively enriched for memory T-cells, while naive T-cells, which have the potential to induce graft-versus-host disease, are depleted [11]. CD45RA depleted cells are a source of memory cells that contain specific virus T-cells and therefore constitute a very interesting alternative for the treatment of resistant infections [12].

Subject and methods

A male patient in his 70s was admitted in May 2017 with persistent fever (38.5°C) and general poor condition after a recent diagnosis of rheumatoid arthritis treated with prednisone (10 mg/day) and methotrexate (15 mg/week) for 4 months. His past medical history included a mild renal insufficiency.

After a diagnosis of CMV disease (anti-CMV-immunoglobulin [Ig]G positive, CMV viral load of 758 IU/ml-lg 2.9); valganciclovir (900 mg twice daily) was started. 1 week later, the patient was asymptomatic and CMV viral load was non-detectable. After 4 weeks of treatment, the lymphocyte and neutrophil counts remained stable, valganciclovir was discontinued and the patient developed a urinary sepsis.

Chronic lymphocytic leukaemia B was then diagnosed. Based on hypogammaglobulinaemia (0.36 g/dl [normal value (NV): 0.7 to 1.3]) and recurrent infections, treatment with bendamustine was started (lymphocytes 3,350/μl, CD4 465/μl [NV: 370 to 1,468]; CD8 282/μl [NV: 183 to 799]). 1 week later, lymphocyte count was 400/μl and the patient was admitted for pneumonia.

1 week after the 2nd cycle of bendamustine, the patient's general condition deteriorated, revealing severe lymphopenia (100/μl) and a plasma level of CMV DNA of 1,170 IU/ml-lg 3.1. Valganciclovir was restarted and 2 weeks later, CMV viral load was non-detectable, and after 4 weeks of treatment, valganciclovir was given prophylactically. After ending bendamustine, lymphopenia persisted (400/μl) with a CD4⁺ and CD8⁺ count of 14/μl and 10/μl, respectively.

When valganciclovir was withdrawn, CMV replication was again detected 2 weeks later and followed by CMV syndrome and end-organ disease (pneumonitis and colitis). Valganciclovir was restarted and the viral load returned to non-detectable levels with symptomatic remission. The use of foscarnet or cidofovir was excluded due to moderate renal failure and newer anti-virals as maribavir, letermovir or brincidofovir were not available. Leflunomide was added to valganciclovir and stopped after 2 weeks due to hepatotoxicity. The patient had recurrent and severe life-threatening bacterial infections (severe recurrent pneumonias and sepsis). Given the absence alternatives and the important clinical deterioration of the patient, adoptive immunotherapy was proposed (Figure 1).

OPEN IN VIEWER

Figure 1

Clinical and laboratory findings

Patient and donor gave written informed consent, and adoptive T-cell transfer was performed in accordance with the regulations of the institutional ethics committee.

Results

Patient

Valganciclovir was stopped 1 week before beginning adoptive T-cell transfer to facilitate viral replication and enhance the antigenic stimulation of the infused T-cells. Adoptive immunotherapy with infusion of CD45RA-depleted donor cells was administered intravenously, once weekly, in the day hospital. The first dose (25,000 cells/kg) was infused immediately after depletion. This was followed by 4 infusions with 50,000 cells/kg and by 13 infusions with 100,000 cells/kg. The doses are based on the total number of depleted CD45RA cells. A total of 18 infusions (99,125,000 cells), were administered. The cells were infused immediately after thawing. No adverse events were observed and safety laboratory values remained stable.

Plasma level of CMV DNA and cell count responses during and after adoptive T-cell transfer

After initiating the infusion of CD45RA-depleted T-cells, no viral replication was detected (Figure 1B).

Lymphocytes, CD4⁺ and CD8⁺ cells were counted weekly. Since the first infusion, progressive increases in cells counts were observed. A normal value of lymphocytes was recovered after six infusions, and of CD8⁺ and CD4⁺ cells after five and nine infusions, respectively. After 12 months, all these cell counts remained in the normal range (Figure 1C).

CMV-Specific T-Cell Responses at the End of Adoptive T-cell Transfer

Lymphocyte subpopulations before and after 16 infusions showed an increment in both effector CD4⁺ (3.2% versus 8.4%) and CD8⁺ (20.9% versus 67.2%; Figure 2A). After treatment, both CD4⁺ and CD8⁺ contained CMV-specific cells (1.08% of CD4⁺IFNγ+ cells and 3.06% of CD8⁺IFNγ+ cells; Figure 2B).

OPEN IN VIEWER

Figure 2

Immunological studies

Clinical outcomes and long-time evolution after adoptive T-cell transfer

1 year after adoptive immunotherapy, the patient remained asymptomatic and had not presented any infectious disease. CMV viral load remained undetectable (Figure 1B). CD4⁺ and CD8⁺ counts persisted in the normal range (Figure 1C) and CMV-specific T-cells were detected (CD4⁺IFNγ+: 0.4%, CD8⁺IFNγ+: 4.39%; Figure 2A).

Discussion

We describe the use of adoptive immunotherapy as a new therapeutic approach for the treatment of CMV disease in an immunocompromised non-transplant patient with cytoxicity related to standard antiviral treatment.

This therapy is based on the crucial role of cellular immunity in protecting against CMV disease. There are several strategies to obtain CMV-specific T-cells but some of them, based on the transfer of CD8⁺ and CD4⁺ cells, are specially interesting because despite the importance of a rapid specific CD8⁺ response, immune restitution of CD4⁺ T-lymphocytes is needed in the recall responses to latent infections to help cytotoxic T-cells [13].

CliniMACS Cytokine Capture System is a method that allows a rapid enrichment of CMV pp65-specific CD4⁺ and CD8⁺ IFN-γ secreting cells [14]. For this strategy, we need a CMV seropositive donor with a sufficient number of peptide-specific T-cells in peripheral blood and a certain HLA compatibility is needed, so that donor's T lymphocytes can recognize the virus antigen presented by the receptor cells (we decided to select an haploidentical donor). In our study, we did not have a donor that accomplished the requirements. In these cases, the use of CD45RA-depleted T-cells is an alternative as they represent a source of memory cells that contain virus-specific T-cells [15].

At the moment, there is no experience in the infusion of specific CD45RA-depleted T-cells except in patients undergoing haematopoietic stem cell transplantation. In these patients, cell transfer enhances immune reconstitution and is associated with a lower rate of infections. The presence of few naive T-cells but abundant memory T-cells preserves memory-immunity and decreases the incidence of graft-versus-host disease [12].

In our patient, initial cytotoxicity was induced by bendamustine. Valganciclovir inhibits viral replication but it plays an important role hindering cell recovery. No resistance to valganciclovir was identified although its suppression was followed by early viral replication and CMV disease, but not by cellular recovery.

In previous studies, the dose of infused cells depleted of CD45RA varies from 25,000 cells/kg [16] to 528x10 ⁶ cells/kg [17]. Repeated cell infusion with dose escalation is safe [18]; therefore, we decided to start with a lower number of cells (25,000/kg) and increase the dose in repeated infusions up to a total of 99,125,000 cells. Further studies are required to determine the optimal timing and T-cell dose.

The primary end point was to control CMV through the infusion of CD45RA-depleted T-cells without valganciclovir. Secondary end points were the cell recovery (lymphocyte, CD4⁺ and CD8⁺), and the control of recurrent infections. All end points were achieved and maintained after 12-month follow-up with a dramatic and persistent improvement in quality of life and without adverse events.

Additionally, we observed a clear increase in effector CD8⁺ cells and a specific cellular response to CMV after treatment and maintained after 1-year follow-up. This response may be responsible for the control of CMV disease.

We do not know if the patient had specific CMV cells prior to infusions of CD45RA depleted cells because it was not possible to isolate enough cells to do the analysis. However, it is possible that he had not a specific CMV immune response because he had CMV disease reactivation when the antiviral was stopped suggesting that the patient did not have specific immune response.

We attempted to know the mechanism implied in the specific CMV response and in the immune system recovery. It could be that the specific CMV cells detected in patient peripheral blood after several infusions were donor cells, but we have not detected them with either studies of chimerism or by using PCR-SSP for HLA typing (data not shown). It could be because the techniques might not be sensitive enough to detect the cells infused in peripheral blood or because the donor cells have been able to migrate to peripheral tissues such as the lung or intestine where the CMV is located to lyse the infected cells and they aren't in peripheral blood. Then, at least the specific CMV cells detected in peripheral blood were from the patient.

The patient immune system recovery could be a critical point. The infused cells could be an allogenic stimulus that activate the patient immune system via host versus graft reaction, or donor cells could control the CMV infection and this could allow us to keep the patient without antiviral treatment. We have not detected antibodies specific against donor HLA. Therefore, it is possible that the withdrawal of treatment could allowed the patient's immune system to recover from the toxicity produced by both antivirals and chemotherapy, so that it is ultimately the patient's own immune system that was currently controlling the infections. Therefore, the depleted cells of CD45RA have contributed to the patient immune system recovery.

In conclusion, adoptive T-cell therapy with CD45RA-depleted donor cells can be considered a salvage therapy for the treatment of CMV disease in immunocompromised non-transplant patients when ganciclovir is not recommended.