Vaccination with peptide-loaded dendritic cells (DCs) has been shown to be potent immunostimulatory therapy for the management of serious infections. After allogeneic stem cell transplant (SCT), a prolonged and severe immune deficiency often leads to infectious complications. Human cytomegalovirus (HCMV) infection is one such life-threatening complication after allogeneic SCT. A phase 1/2 study including 24 allogeneic SCT recipients at high risk for HCMV disease was performed to analyze the feasibility and efficacy of vaccination with HCMV peptide—loaded DCs. No acute adverse effects were observed, and a significant clinical benefit could be demonstrated in comparison to our historical control group. An induction or expansion of HCMV-specific cytotoxic T lymphocytes was observed in 5 patients after DC vaccination.
The principle of allogeneic stem cell transplantation (SCT) for malignant diseases aims at 2 different issues: first to replace an irreversibly damaged stem cell compartment with healthy hematopoiesis and second to establish an intact immune system to prevent relapse caused by the graft-versus-tumor effect. However, strong immunosuppression is absolutely essential for the prevention severe graft-versus-host disease (GVHD), but it leads to infectious complications.
Dendritic cell (DC) vaccination is a pathogen specific and highly immunostimulatory approach to fighting different infections and malignancies , but it has never been evaluated in the setting of allogeneic SCT before. The risk for GVHD still makes clinicians wary of immunostimulatory therapies such as DC vaccination. Thus, we performed a clinical trial to evaluate DC vaccination in patients after allogeneic SCT. We chose human cytomegalovirus (HCMV) as the target for our DC vaccination. HCMV infection not only continues to be one of the most important and life-threatening complications after allogeneic SCT [2, 3] but has also been shown to be a target well suited for cellular immunity . In addition, reconstitution of HCMV-specific immune responses after SCT, either endogenously or by transfer of HCMV-specific T cells from the stem cell donor, has been documented to prevent HCMV disease [4–8]. According to established protocols for the generation of sufficient numbers of donor-derived HCMV-specific T cells for adoptive immunotherapy, the donor has to be HCMV seropositive; in addition, up to 8 weeks of in vitro culture are required to deplete alloreactive T cells and to obtain enough HCMV-specific T cells for successful transfer . Thus, alternative strategies for inducing HCMV-specific T cell responses are highly warranted.
Therefore, we treated HCMV-seropositive patients with donor-derived DCs pulsed with HCMV peptides. All patients were at high risk for developing HCMV disease (HCMV-seropositive patient and HCMV-seronegative donor and/or receipt of a T cell—depleted transplant) [9, 10].
Because this was, to our knowledge, the first DC vaccination trial in allogeneic SCT recipients, the primary objective was safety, specifically with regard to vaccination-related toxicities and the induction or aggravation of GVHD. The secondary objective was the efficacy of DC vaccination. Because all patients with active HCMV infection received antiviral chemotherapy [3, 11] at the time of DC vaccination, short-term control of viral load was not a suitable efficacy parameter. Thus, reconstitution of HCMV-specific cytotoxic T lymphocyte (CTL) responses and long-term control of HCMV infection were used as parameters for vaccination efficacy.
Patients, materials, and methods. In this single-armphase 1/2 trial, 24 patients at high risk for HCMV infection with different underlying diseases were included (see table 1) between March 2001 and February 2004.
Patients at high risk for HCMV infection were defined as HCMV-seropositive SCT recipients who received either a transplant from an HCMV-seronegative donor or a T cell—depleted graft. These patients were included during active HCMV infection (therapeutic vaccination) and before or after HCMV reactivation (prophylactic vaccination). An additional 17 patients (who received transplants between January 2001 and December 2004) met these high-risk criteria but could not be vaccinated because of an HLA type for which no HCMV peptides are defined or an inaccessibility of donor samples to generate the vaccine. We used these 17 patients as our historical untreated control group.
Except one, all stem cell grafts were T cell depleted, either in vitro (CD34+ cell selection) or in vivo (ATG [Fresenius] or Alemtuzumab [MSO-Pharma]). All patients were HCMV seropositive and received prophylactic (n = 5) or therapeutic (n = 19) vaccination in addition to routinely administered antiviral therapy, as described elsewhere [3, 11]. Nineteen stem cell donors were HCMV seronegative, and 5 were HCMV seropositive. The study protocol was approved by the local ethical committee, and all patients and donors gave written, informed consent.
For DC generation, 200–400 mL of peripheral blood was drawn from the stem cell donor. In 2 patients, no blood could be obtained from the unrelated donor. Thus, DCs were generated from donor-derived peripheral blood mononuclear cells (PBMCs) obtained from the patients after transplant.
DCs were generated under good manufacturing practice conditions as described elsewhere [1, 12], with minor modifications. Because a certified monocyte isolation reagent was available from the end of 2003 on, we changed the isolation procedure of monocytes at that time. Plastic adherence was replaced by the positive selection of CD14+ cells (patients 17–24). For positive selection, PBMCs were incubated with CD14 reagent (Miltenyi Biotech) and selected using the Clinimacs instrument in accordance with the instruction protocol. After magnetic selection, monocytes were seeded (3 × 106 cells/3mL/well) into 6-well plates and cultured in CellgroDC (Cellgenix) medium supplemented with 100 ng/mL human recombinant granulocyte macrophage colony-stimulating factor (Leukine; Berlex), 1000 IU/mL interleukin-4, and 10 ng/mL tumor necrosis factor—α (both from R&D Systems). DC cultures were fed with fresh medium and cytokines every other day, and cell differentiation was monitored by light microscopy. The expression of cell-surface molecules on DCs was analyzed by flow cytometry after 10 days of culture.
Cells displayed a morphology typical of DCs by light microscopy and expressed the typical surface markers of mature DCs (CD1a+CD14−CD40+CD80+CD83+CD86+). After the cells were harvested, DCs were incubated with 50 µg/mL of HLA class I peptides from the HCMV proteins pp65 and pp150, which had been identified previously  (for sequences, see table 1). To minimize the risk for vaccination-induced GVHD being mediated by residual T cells in the vaccine, DCs were irradiated with 30 Gy. Final sterility controls were performed, and 1 × 105−5 × 106 DCs/m2 were administered subcutaneously near the inguinal lymph nodes. To assess the adverse effects of the treatment, all patients had to undergo the following examinations before treatment: physical examination and determinations of Karnofsky score, body weight, body size, body surface, clinical grade of GVHD, serum chemistry, plasmatic coagulation, and differential blood count. For the evaluation of long-term adverse effects and long-term control of HCMV infection, we defined a monitoring period of at least 3 months after vaccination. HCMV diagnostics were performed weekly using polymerase chain reaction (PCR) and pp65 antigenemia assays, as described elsewhere [14, 15], with modifications of the PCR method. HCMV-specific T cell responses were analyzed before vaccination, twice during the first month after vaccination, and thereafter monthly 5 times by flow cytometry using tetrameric HLA/peptide complexes and/or intracellular interferon-γ staining.
Results and discussion. Allogeneic SCT is associated with a long-standing and severe immune deficiency leading to a high risk for opportunistic infections. Therefore, different strategies to improve immune reconstitution after transplant have been evaluated in the setting of allogeneic SCT. DC vaccination is one of the strategies that allows the boosting of immune reconstitution or the induction of primary immune responses to infectious pathogens and might thus help prevent or treat infectious complications after SCT.
In the present study, DC vaccination was evaluated for the first time after allogeneic SCT in a phase 1/2 trial. Because HCMV still causes high morbidity and mortality after SCT and because HCMV proteins are known to be strongly immunogenic, we postulated that patients at high risk for HCMV disease after allogeneic SCT might benefit from vaccination withHCMV peptide—pulsed DCs.
During the monitoring period after DC vaccination, 6 patients died of reasons not related to the vaccination or HCMV infection. Insufficient follow-up samples were available from 1 patient. Therefore, 24 patients were evaluable for acute adverse effects, and 17 patients were evaluable (table 1) for long-term adverse effects and long-term control of HCMV infection.
No acute adverse effects occurred in any of the 24 patients. Neither systemic reactions, such as cardiovascular dysfunction, nor local irritations of the skin or of the subcutaneous tissue were observed. Two months after vaccination, 1 patient (patient 12) developed acute grade III GVHD of the skin and gut. The GVHD reaction was controlled by corticosteroid therapy and resolved completely. Because of a time lag of 2 months between vaccination and the onset of GVHD and a subtherapeutic level of cyclosporin A before the onset of acute GVHD, a causative relationship between DC vaccination and GVHD is unlikely. Former trials investigated DC vaccination only in an autologous setting . Thus, very thorough monitoring of the induction or aggravation of alloreactivity in this first study of allograft recipients was performed. The lack of induction or aggravation of acute and chronic GVHD in this trial—although in a limited number of allograft recipients—indicates that the transfer of donor-derived DCs pulsed with HCMV peptides does not stimulate or expand alloreactive T cells. No long-term adverse effects of DC vaccination were observed in any of the 17 evaluable patients. Taken together, this phase 1/2 study indicates that vaccination with donor-derived DCs can be performed safely in allogeneic SCT recipients, even in those receiving haploidentical SCT (table 1).
This trial was designed to evaluate the safety of DC vaccination in allogeneic SCT recipients. Because the possible risks, especially the induction of GVHD, could not be predicted, we chose a patient cohort that might benefit from this experimental treatment: patients who had received long-term antiviral chemotherapy for HCMV infection, those with chemotherapy-refractory HCMV infection, and HCMV-seropositive recipients with HCMV-seronegative donors. All patients included were at a very high risk of developing HCMV disease, which is still associated with high mortality. As a secondary objective, the treatment efficacy was analyzed. Because most of the patients had uncontrollable HCMV infection when included in the study, they received antiviral therapy during the vaccination procedure and the short-term follow-up period. Thus, to assess the efficacy of DC vaccination, we studied the long-termcontrol of HCMV infection as well as the induction and expansion of HCMV-specific T cell responses. Seventeen patients (6 receiving prophylactic vaccination and 11 receiving therapeutic vaccination) were evaluable for efficacy.
Seven of the 17 evaluable patients showed an increased (2 patients had already low numbers of circulating HCMV-specific CTLs before vaccination) or a de novo induction (n = 5) of HCMV-specific CTL responses a mean of 23 days (range, 8–35 days) after DC vaccination. In total, DC vaccination induced or expanded the HCMV-specific CTL response in 7 (41%) of 17 evaluable patients. Although the vaccination procedure induced T cell responses and seemed to provide a clinical benefit in most vaccinated patients, we could not ascertain a significant relationship between the occurrence of a HCMV-specific T cell response and the long-term control of HCMV infection (P > .05; Fisher's exact test).
When we analyzed the prophylactically vaccinated population, all patients showed long-termcontrol of HCMV infection. Among these patients, 3 of 6 had detectable HCMV-specific CTL responses occurring (n = 2) or maintained (n = 1) after DC vaccination. Only 1 of the prophylactically treated patients who did not develop a detectable HCMV-specific CTL response (patient 16) had transient HCMV reactivations without antiviral therapy.
In the therapeutically vaccinated group, HCMV reactivation resolved completely in 9 (82%) of 11 patients. Five of these 9 patients showed a boost in the HCMV-specific CD8+ cell response after DC vaccination. The other 4 patients controlled their HCMV infection despite lacking a specific T cell response. The remaining 2 patients undergoing therapeutic vaccination (patients 6 and 11) did not develop an HCMV-specific CTL response. Both had ongoing HCMV viremia and developed nonfatal HCMV disease.
Among the successfully treated patients, one (patient 5; see figure 1) received DC vaccination for HCMV infection with a virus strain cross-resistant to ganciclovir, foscarnet, and cidofovir. This patient cleared HCMV 14 days after vaccination and developed a functional HCMV-specific CTL population. However, after high-dose steroid therapy for preexisting severe autoimmune hemolysis, HCMV infection recurred following the disappearance of the functional circulating HCMV-specific CD8+ CTL population. After a second DC vaccination, this patient definitely cleared HCMV infection. With regard to the multiresistant HCMV strain, this case clearly indicates the therapeutic potential of DC vaccination. Thus, we postulate that the induction and expansion of HCMV-specific CTLs as well as the improved long-term control of HCMV as observed in 5 patients is due to an improved presentation of HCMV epitopes by the transferred DCs.
Only 2 patients (patients 10 and 13) could be examined for CD4+ cell responses before and after DC vaccination because of the limited availability of cells from patients with lymphopenia. Neither patient 10, who developed a vaccine-specific CD8+ cell response, nor patient 13, who had no response to the vaccine peptide, had HCMV-specific CD4+ cells. These limited data might indicate that DCs pulsed with HLA class I motif peptides are able to induce and expand CTL responses against HCMV in the absence of support from specific CD4+ T cells.
In conclusion, vaccination with HCMV peptide—pulsed DCs is a safe and feasible method to use after allogeneic SCT, even in patients who have received haploidentical SCT (n = 2). Our results show DC vaccination to induce (n = 5) or improve (n = 2) protective peptide-specific T cell responses in allogeneic SCT recipients without relevant adverse effects.
This effect was not limited to patients receiving stem cells from HCMV-seropositive donors, which is essential when performing adoptive immunotherapy with donor-derived HCMV-specific T cells. In fact, DC vaccination was able to induce a CD8+ cell response even in recipients of a transplant from a HCMV-seronegative donor without any detectable HCMV-specific CD8+ cell precursor frequencies (figure 1A). This represents an additional strategy for HCMV-specific immunotherapy, because the selection for immunotherapy would no longer depend on the HCMV serostatus of the donor.
To assess the overall efficacy of our vaccination procedure, we compared the long-term control of infection in 17 evaluable study patients with that in a historical untreated control group of 17 patients who were also at high risk for HCMVreactivation (HCMV-seropositive patient/HCMV-seronegative donor) but who could not be included in the vaccination trial for reasons described above. Here, we were able to show a significant benefit for the patients who received DC vaccination (P < .01, Fisher's exact test): only 2 (11.8%) of 17 vaccinated patients had a reactivation of HCMV disease after DC application, whereas 12 (70.6%) of 17 patients in the control group had a reactivation of HCMV during the observation period of at least 3 months. We cannot explain the discrepancy between the missing proof of T cell responses after vaccination in some patients and the clinical benefit in the study population, compared with that in the control group. But this might be due to an induced immune response below the detection limits of our analytical assays.
Because no relevant adverse effects were observed in this first DC vaccination trial in allogeneic SCT recipients and because induction and expansion of HCMV-specific T cell responses could be demonstrated in 41% of the patients, further investigations will evaluate the efficacy of DC vaccination in a larger cohort of patients.