Dynamics of Sendai Virus Spread, Clearance, and Immunotherapeutic Efficacy after Hematopoietic Cell Transplant Imaged Noninvasively in Mice

ABSTRACT

There are no approved vaccines or virus-specific treatments for human parainfluenza viruses (HPIVs), which have recently been reclassified into the species Human respirovirus 1, Human respirovirus 3, Human rubulavirus 2, and Human rubulavirus 4. These viruses cause morbidity and mortality in immunocompromised patients, including those undergoing hematopoietic cell transplant (HCT). No small-animal models for noninvasive imaging of respiratory virus infection in the HCT host exist, despite the utility that such a system would offer to monitor prolonged infection, its clearance, and treatment options. We used a luciferase-expressing reporter virus to noninvasively image in mice the infection of murine respirovirus (strain Sendai virus [SeV]), the murine counterpart of HPIV1. Independent of disease severity, the clearance of infection began approximately 21 days after HCT, largely due to the recovery of CD8⁺ T cells. Immunotherapy with granulocyte colony-stimulating factor (G-CSF) and adoptive transfer of natural killer (NK) cells provided a limited therapeutic benefit. Treatment with a fusion (F) protein-specific monoclonal antibody arrested the spread of lung infection and reduced the disease severity even when treatment was delayed to up to 10 days postinfection but had little observable effect on upper respiratory tract infection. Adoptive transfer of virus-specific T cells at 10 days postinfection accelerated the clearance by 5 days, reduced the extent of infection throughout the respiratory tract, and reduced the disease severity. Overall, the results support investigation of the clinical treatment of respiratory virus infection in the HCT host with monoclonal antibodies and adoptive T-cell transfer; the imaging system should be extendable to other respiratory viruses, such as respiratory syncytial virus and influenza virus.

IMPORTANCE Parainfluenza viruses are a major cause of disease and death due to respiratory virus infection in the immunocompromised host, including those undergoing bone marrow transplantation. There are currently no effective treatment measures. We noninvasively imaged mice that were undergoing a bone marrow transplant and infected with Sendai virus, a murine parainfluenza virus (respirovirus). For the first time, we show the therapeutic windows of adoptive T-cell therapy and treatment with a monoclonal antibody to the fusion (F) protein in clearing Sendai virus from the respiratory tract and reducing disease severity. Mice tolerated these treatments without any detectable toxicity. These findings pave the way for studies assessing the safety of T-cell therapy against parainfluenza virus in humans. Adoptive T-cell therapy against other blood-borne viruses in humans has been shown to be safe and effective. Our model of noninvasive imaging in mice that had undergone a bone marrow transplant may be well suited to track other respiratory virus infections and develop novel preventive and therapeutic strategies.

INTRODUCTION

Allogeneic hematopoietic cell transplant (HCT) is a curative modality for patients with certain hematologic malignant neoplasms and nonmalignant disorders. Posttransplant infections contribute significantly to the morbidity and mortality arising from the procedure (1, 2). The ablation of their immune system by the preparative regimen renders HCT recipients susceptible to infections while they await immune reconstitution from donor stem cells. The process of determining the kinetics of immune cell reconstitution after HCT and identifying the cell types needed to control specific pathogens provides opportunities to dissect host-pathogen interactions and devise novel immunotherapeutic strategies for use in immunocompromised and perhaps even immunocompetent patients. HCT in a mouse model offers a unique platform to understand the impact of novel therapies in an engineered host immune system.

Respiratory syncytial virus (RSV), influenza virus, human parainfluenza viruses (HPIVs), and human metapneumovirus (hMPV) are the most common viral causes of respiratory tract infections in HCT recipients (3). Over 30,000 allogeneic transplants are performed annually worldwide (4). The present study focused on parainfluenza virus (PIV) infection, which is a leading cause of morbidity and mortality after HCT (5). HPIV1, HPIV2, and HPIV3 infect 5% to 7% of HCT recipients. Of these, 24% to 50% develop pneumonia, which is fatal in up to 75% of cases (5–11). Currently, there are no effective strategies to prevent or treat HPIV infections. Mortality has been correlated with the progression of infection to the lower respiratory tract (LRT) (10). Consequently, early intervention shortly after the diagnosis of PIV upper respiratory tract infection may improve outcomes and reduce mortality. The receipt of an HCT from an unrelated donor and the consequent development of graft-versus-host disease (GVHD) are associated with the progression of PIV infection to the LRT (5). Infection within the first 100 days after the transplant also contributes to the risk of progression. The human parainfluenza viruses have recently been renamed to the species Human respirovirus 1 (strain HPIV1), Human rubulavirus 2 (strain HPIV2), Human respirovirus 3 (strain HPIV3), and Human rubulavirus 4 (strain HPIV4).

There is a fundamental gap in our understanding of respiratory virus infection and the control thereof in HCT recipients due to a scarcity of suitable small-animal models and preclinical studies on infection clearance and candidate therapeutics. To understand the immune responses to viral infection, many studies have relied on models involving mice that are genetically deficient in their immune responses (e.g., severe combined immune deficient mice or muMT knockout mice) (12, 13) or are immunosuppressed by drug treatment (e.g., with steroids or cyclophosphamide) (14, 15). The overwhelming majority of studies on infection in HCT recipient mice have focused on systemic DNA viruses, such as cytomegalovirus (CMV), Epstein-Barr virus (EBV), or lymphocytic choriomeningitis virus (LCMV) (16–20). These studies have provided insights into viral pathogenesis and the effect of engraftment on infection, and they have validated cellular immunotherapy as an antiviral treatment in HCT recipients.

There have been few published studies on respiratory RNA virus infection in small-animal models of HCT. With respect to influenza A/Puerto Rico/8/34 (H1N1) virus infection in mice that received syngeneic bone marrow transplants (BMTs), CD4⁺ and CD8⁺ T cells have been associated with protection (21, 22), and interleukin-1 (IL-1) has shown therapeutic potential (23). Sendai virus (SeV), a member of the Respirovirus genus of the family Paramyxoviridae, is the murine counterpart of HPIV1 (24). The virus strain SeV currently has the species name Murine respirovirus. SeV and HPIV1 are similar genetically and antigenically but counteract innate immune responses in a species-specific manner. Thus, a mouse model of natural SeV infection provides a useful model of PIV infection (24–26). Previously, impaired CD8⁺ T-cell function was linked to prolonged and more severe SeV infection in mice that had received allogeneic BMTs (27). The goals of this work were to develop and use allogeneic models of SeV infection in mice undergoing HCT to analyze the kinetics of respiratory virus spread and clearance and to investigate immunotherapeutic strategies.

RESULTS

Noninvasive imaging of mild and severe infection after HCT.

To recapitulate the clinical associations of respirovirus disease severity in the HCT host (a host receiving total irradiation, allogeneic HCT, and early infection [5]), we irradiated BALB/cJ mice, 1 day later we intranasally inoculated them with the recombinant SeV (rSeV)-luciferase(M-F*) reporter virus, and at 6 h postinfection (hpi) we transplanted a T-cell-depleted bone marrow graft from C57BL/6J donor mice. The rSeV-luciferase(M-F*) reporter virus was a nonattenuated, recombinant SeV that expresses firefly luciferase in the cytoplasm of infected cells, allowing noninvasive daily imaging of viral infection in living mice (28–30). The use of bone marrow from fully major histocompatibility complex (MHC)-mismatched mice (C57BL/6J mice → BALB/cJ mice) for BMTs, an approach commonly used to study GVHD (31), provided an added risk factor for severe infection. Irradiation eliminated peripheral neutrophils and lymphocytes within 3 days, and these cells began to increase by days 13 and 21 posttransplant, respectively (Fig. 1A and B).

FIG 1.

Mild SeV infection in transplant recipients. BALB/cJ mice were irradiated, infected with 7,000 PFU of SeV in a 5-μl volume, and for transplantation given a T-cell-depleted bone marrow graft from C57BL/6J mice. (A and B) Peripheral blood neutrophils (A) and lymphocytes (B) were measured periodically in control transplant recipient mice, transplant recipients infected with SeV, and healthy control mice infected with SeV. (C) Percent change in body weight relative to the weight at the start of the experiment. (D to F) Bioluminescence emitted from the tissues of living mice after i.p. injection of 150 mg/kg d-luciferin. Zonal values are shown for the nasopharynx (N) (D), trachea (T) (E), and lungs (L) (F). Data are representative of those from 2 independent experiments with 5 mice per group in each experiment. Error bars represent standard deviations.

To mimic the infection dynamics seen after contact transmission and to induce disease of mild severity (32), we intranasally inoculated mice with 7,000 PFU of SeV in a 5-μl volume. For both mice that underwent HCT and healthy control mice, infection initiated in the nasopharynx and spread to the trachea and lungs 1 day later (Fig. 1D to F). Nasal infection in immunocompetent mice peaked at days 2 to 4 (approximately 10⁸ photons/s) and cleared by day 9 (Fig. 1D). Nasal infection in mice that underwent HCT peaked at levels (10^8.5 photons/s; P < 0.001) higher than those in immunocompetent mice and remained above 10⁸ photons/s for nearly 3 weeks. Lung infection in immunocompetent mice peaked on days 4 to 5 (approximately 10^6.2 photons/s) and cleared by day 7, while lung infection in mice that underwent HCT progressed to a significantly higher peak level (10^6.8 photons/s; P < 0.02) at days 15 to 17 and started to clear after day 21 (Fig. 1F). Even though the transplant recipients had greater lung bioluminescence over a longer period than did control mice, contributing to a delayed recovery of weight (Fig. 1C), the weight loss in mice that underwent HCT was typically no more than 10% during recovery, and the rate of survival was 100% (Fig. 2E).

FIG 2.

Severity of SeV infection in transplant recipients modulated by the inoculated dose and volume. BALB/cJ mice were irradiated, infected with SeV (in various doses and volumes), and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. Differential inoculation yielded infection that was mild (with 7,000 PFU SeV in 5 μl), moderate (with 700 PFU SeV in 30 μl), or severe (with 7,000 PFU in 30 μl). (A to C) Bioluminescence in the nasopharynx (A), trachea (B), and lungs (C); (D and E) clinical signs in terms of the percent change in starting weight (D) and survival (E); (F) lymphocyte counts in peripheral blood. The error bars represent standard deviations. Data are representative of those from 2 or more experiments with 5 mice per group in each experiment.

To induce moderate and severe infections, we intranasally inoculated mice with 30 μl of SeV at dosages of 700 and 7,000 PFU, respectively. Compared to the 5-μl inoculation, which yielded a peak lung bioluminescence of <10⁷ photon/s, a 30-μl inoculation increased the lung infection to 10^7.9 and 10^8.5 photon/s for the 700- and 7,000-PFU doses, respectively (P < 0.05) (Fig. 2C). Regardless of the dose or the volume inoculated, clearance of the lung infection began after day 21 (Fig. 2C), and nasal and tracheal infections were similar in magnitude and kinetics (Fig. 2A and B). Transplant recipients inoculated with 7,000 PFU in 30 μl suffered 100% mortality after losing over 25% of their body weight, while 100% of the mice in the other groups survived (Fig. 2D).

Posttransplant lymphocyte recovery and viral clearance.

Both lymphocyte recovery and viral clearance began approximately 21 days posttransplant independently of disease severity (Fig. 1 and 2). To determine the relative contributions of lymphocyte subsets to clearance, we inoculated BALB/cJ mice with 7,000 PFU of SeV in 5 μl and collected peripheral blood at the times of peak (day 21) and cleared (day 27) infection. B-cell (B220⁺) and NK-cell (CD49b⁺) chimerism was approximately 90% or higher at both time points, while T-cell chimerism was substantially lower (Table 1). Chimerism is the extent of engraftment, which is defined as the percentage of a cell population from the donor after HCT. At the time of peak infection, lymphocytes consisted of 54% B cells, 33% CD4⁺ T cells, and less than 5% each CD8⁺ T cells and NK cells (Fig. 3A). After clearance on day 27, B-cell levels decreased, CD4⁺ T-cell levels remained almost unchanged, NK-cell levels increased, and CD8⁺ T-cell levels increased slightly. In immunocompetent mice, the proportions of lymphocytes near the time of peak infection and after clearance remained relatively constant, with approximately 50% B cells, 30% CD4⁺ T cells, 10% CD8⁺ T cells, and 10% NK cells (30). Thus, the lymphocyte proportions measured here in the transplant recipients were similar at the time of peak infection for B and CD4⁺ T cells but reduced for NK and CD8⁺ T cells. Conversely, by the time that the infection had cleared, the proportion of B cells in the transplant recipients, which were predominantly donor B cells, was approximately 2.5-fold lower than that in immunocompetent mice, whereas the proportion of NK cells, which were predominantly donor NK cells, was approximately 3-fold higher. Serum antibody responses on days 28 and 41 in mice that underwent HCT and that were infected with SeV were significantly higher than those in uninfected control mice that underwent HCT (P < 0.001) yet were significantly lower than those in SeV-infected immunocompetent mice (P < 0.001) (Fig. 3B). Thus, the B-cell responses remained functional, albeit at a reduced level.

TABLE 1.

Chimerism of transplant recipient mice infected with SeV^a

Day	Chimerism (% of donor cells)
	CD4+	CD8+	B220+	CD49b+
21	1.9 ± 1.7	15.1 ± 9.9	88.4 ± 5.4	97.2 ± 0.8
27	29.5 ± 17	37.2 ± 20.6	88.2 ± 1.9	94.2 ± 1.8

^aBALB/cJ mice were irradiated, infected, and for transplantation received a T-cell-depleted bone marrow graft derived from C57BL/6J mice. Blood was collected and analyzed by flow cytometry for chimerism by using the H2Kb and H2Kd markers. There were 5 mice per time point.

FIG 3.

T-cell and B-cell responses and clearance of infection in the lungs. BALB/cJ mice were irradiated, infected with SeV, and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. (A) Flow cytometry analysis of peripheral blood cells on day 21 (D21) and day 27 (D27) after transplantation. The percentages of CD4⁺ T cells (black), CD8⁺ T cells (blue), B cells (orange), and NK cells (green) are shown. (B) Serum IgG values from peripheral blood of SeV-infected mice at 28 or 41 days after the transplant as measured by ELISA. Values of the optical density at 405 nm (OD 405) are shown for control mice that did not undergo transplantation (black), uninfected transplant recipients (orange), and SeV-infected transplant recipients (red). For both panels A and B, mice were infected with 7,000 PFU of SeV in 5 μl. Error bars represent the standard deviations, and there were 5 mice per time point. (C) Histopathology of lung sections at 15 (top), 20 (middle), and 26 (bottom) days after the transplant for mice infected with 7,000 PFU of SeV in 30 μl. T cells were stained with anti-CD3 antibody, and SeV was detected with an SeV-specific antibody. There were 2 mice per time point. H&E, hematoxylin and eosin.

For histopathology, we inoculated BALB/cJ mice with 7,000 PFU in 30 μl. The level of T-cell infiltration in immunocompetent control mice was relatively low at the time of peak infection (day 5) but robust by the time of clearance (day 8) (data not shown). In mice that underwent HCT, peak infection (day 20) was more extensive than that in control mice and was associated with inflammation, a loss of normal architecture, and extensive T-cell infiltration on day 26, coinciding with clearance (Fig. 3C).

To determine whether B and T cells were needed for clearance, we inoculated 700 PFU of SeV in 30 μl and performed a fully MHC-mismatched transplant using a T-cell-depleted bone marrow graft from C57BL/6J mice depleted of B or T cells. B-cell depletion did not affect SeV infection or clearance in mice that underwent HCT (Fig. 4A to C). T-cell depletion prevented clearance, leading to prolonged infection (Fig. 4A to C), despite a 2-fold increase in the neutrophil count (Fig. 4D), the presence of B cells, increased amounts of NK cells (Fig. 4E), and antibody responses (Fig. 4F) in the bronchoalveolar lavage fluid (BALF). Thus, T lymphocytes played a dominant role in infection clearance in transplant recipients independently of the antibody responses.

FIG 4.

Depletion of T cells, but not B cells, delays SeV clearance after transplantation. BALB/cJ mice were irradiated, infected with 700 PFU of SeV in 30 μl PBS, and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. The mice were treated with anti-CD20 (anti-B cell) at days 13 and 20 or anti-CD3 (anti-T cell) at days 13 and 16 after the transplant. (A to C) Bioluminescence in the nasopharynx (A), trachea (B), and lungs (C); (D) lymphocyte and neutrophil counts in BALF at day 30 after the transplant; (E) flow cytometric analysis of the relative proportions of lymphocyte subsets in BALF at day 30 after the transplant; (F) SeV-specific IgG and IgA levels in BALF determined by ELISA at day 30 after the transplant. The error bars represent the standard deviations, and there were 5 mice per group. P values were determined by Student's t test. *, P < 0.05.

Neutrophil immunotherapy with G-CSF.

Granulocyte colony-stimulating factor (G-CSF) is used clinically posttransplant to shorten the period of neutropenia and reduce infection rates. To determine if an increase in the neutrophil count by G-CSF treatment has therapeutic potential, we irradiated BALB/cJ mice, inoculated them with 7,000 PFU of SeV in 30 μl, performed a fully MHC-mismatched transplant with a T-cell-depleted bone marrow graft from C57BL/6J mice, and treated the recipient mice with 125 μg/kg of body weight G-CSF for 10 days starting on day 5. G-CSF treatment significantly increased the neutrophil counts on days 12 and 20 compared to those in untreated control mice (P < 0.05) (Fig. 5D) but had no apparent effect on infection, morbidity, or mortality (Fig. 5A to C, E, and F).

FIG 5.

Treatment with monoclonal antibody M16, but not G-CSF, reduces SeV infection in the lungs of transplant recipients. BALB/cJ mice were irradiated, infected with 7,000 PFU of SeV in 30 μl PBS, and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. Mice were treated with a monoclonal antibody specific to the SeV fusion protein (M16) 2 days after the transplant, with G-CSF daily, starting at day 5, for 10 days, or with both M16 and G-CSF. (A to C) Bioluminescence in the nasopharynx (A), trachea (B), and lungs (C); (D) neutrophil counts in peripheral blood at 12 or 20 days after the transplant; (E and F) effect of treatments on the percent change in starting weight (E) and survival (F). The error bars represent the standard deviations, and there were 5 mice per group. P values were determined by Student's t test. *, P < 0.05.

Treatment with monoclonal antibody.

To determine the therapeutic potential of virus-specific antibodies in the HCT host, we used an SeV fusion (F) protein-specific monoclonal antibody (M16) (30). We irradiated BALB/cJ mice, inoculated them with 7,000 PFU of SeV in 30 μl, performed a fully MHC-mismatched transplant with a T-cell-depleted bone marrow graft from C57BL/6J mice, and treated the mice with a single dose of M16 at 2 days postinfection (dpi). M16 treatment reduced the lung bioluminescence starting on day 8 (Fig. 5C), reduced the weight loss, and yielded 100% survival (Fig. 5E and F) without affecting nasal or tracheal infection (Fig. 5A and B). G-CSF and M16 combination therapy increased the rates of morbidity and mortality compared to those after M16 monotherapy (Fig. 5E and F), perhaps due to the inflammatory potential of G-CSF or its inhibitory action on NK-cell function (33).

To determine the therapeutic window of M16 in the transplant recipients, we irradiated BALB/cJ mice, inoculated them with 7,000 PFU in 30 μl, performed a fully MHC-mismatched transplant with a T-cell-depleted bone marrow graft from C57BL/6J mice, and treated the mice with M16 shortly before infection (day 0) or at 2, 10, or 14 dpi. Prophylactic administration of M16 antibody substantially reduced the initial and peak infection levels throughout the respiratory tract and accelerated clearance (Fig. 6A to C). M16 treatment at 2 and 10 dpi yielded 100% survival, with treatment at 2 dpi controlling infection better (Fig. 6C and E). M16 treatment at 14 dpi had no effect on the lung signal, but the mice experienced 40% mortality, whereas the untreated SeV-infected transplant recipients experienced 80% mortality (Fig. 6C and E). Overall, the anti-SeV monoclonal antibody effectively controlled the lung infection and reduced the rates of morbidity and mortality in the transplant recipients if it was administered up to 10 days after infection.

FIG 6.

Early versus late therapeutic capacity of M16. BALB/cJ mice were irradiated, infected with 7,000 PFU of SeV in 30 μl PBS, and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. Mice were treated with an anti-SeV monoclonal antibody (M16) specific to the fusion protein shortly before infection (day 0) or at 2 days, 10 days, or 14 days after the transplant. Untreated controls (HCT + SeV) and control mice that did not undergo a transplant (SeV) were included in the study. (A to C) Bioluminescence in the nasopharynx (A), trachea (B), and lungs (C); (D and E) effect of treatment on the percent change in starting weight (D) and survival (E). The error bars represent the standard deviations, and there were 5 mice per group. D, day.

Immunotherapy with NK cells.

We next addressed the therapeutic potential of NK cells, as they engraft early after the transplant (34), can inhibit GVHD, show a graft-versus-leukemia effect (35, 36), and have antiviral properties against cytomegalovirus (37–39). We inoculated BALB/cJ mice with 700 or 7,000 PFU of SeV in 30 μl, performed an HCT using a bone marrow graft from C57BL/6J mice, performed adoptive transfer of NK cells 10 days after the transplant, and administered an IL-15 superagonist (IL-15 SA; IL-15 plus the IL-15 receptor alpha [IL-15Rα]) on days 9, 11, and 13 (40). NK-cell immunotherapy reduced the lung infection significantly (P < 0.05) in mice inoculated with 700 PFU (Fig. 7C) but not those inoculated with 7,000 PFU (Fig. 7G). Combination therapy with NK cells and a tumor antigen-specific monoclonal antibody is currently used to treat cancer through an antibody-dependent cell-mediated cytotoxicity (ADCC) effect (41). However, when we treated the mice that had received a transplant with NK cells and M16 antibody, there was no apparent benefit over treatment with M16 alone (Fig. 7). The adoptive transfer of NK cells on day 14 posttransplant resulted in increased lung infection and weight loss after day 16 and accelerated mortality compared to the time to mortality in untreated mice (Fig. 8C and E).

FIG 7.

NK-cell adoptive transfer partially controls SeV lung infection at a low viral dose after HCT. BALB/cJ mice were irradiated, infected with 700 (A to C) or 7,000 (E to G) PFU of SeV in 30 μl PBS, and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. NK cells were adoptively transferred at day 10, and 3 doses of IL-15 superagonist (IL-15) were injected intraperitoneally at days 9, 11, and 13. Mice were treated with an anti-SeV monoclonal antibody (M16) 2 days after the transplant. The bioluminescence is shown for the nasopharynx (A and E), trachea (B and F), and lungs (C and G). All groups that underwent an HCT were infected with SeV and treated with IL-15. The groups are labeled as untreated (HCT + SeV + IL-15), M16 treated, NK-cell treated, and M16 and NK-cell treated. The error bars represent the standard deviations, and there were 5 mice per group.

FIG 8.

Late NK-cell adoptive transfer increases the SeV signal and the rate of morbidity after the transplant. BALB/cJ mice were irradiated, infected with 7,000 PFU of SeV in 30 μl PBS, and for transplantation given a T-cell-depleted bone marrow graft derived from C57BL/6J mice. NK cells were adoptively transferred at day 14 after the transplant. The IL-15 superagonist (IL-15) was administered intraperitoneally at days 13, 15, and 17. Mice were treated with the M16 anti-SeV monoclonal antibody 2 days after the transplant. (A to C) Bioluminescence measured after i.p. injection of 150 mg/kg d-luciferin and imaging of the nasal cavity (A), trachea (B), and lungs (C) with a Xenogen system; (D) weight loss; (E) survival; (F) flow cytometry of peripheral lymphocytes isolated at day 21 after the transplant. The error bars represent the standard deviations, and there were 5 mice per group.

We next examined the contribution of NK cells to M16-mediated viral clearance using the mouse model of pharmacological immunosuppression, which allowed us to eliminate B-cell functions for an extended time and enhance NK-cell numbers after the discontinuation of the drugs (30). We treated BALB/c mice with dexamethasone and cyclophosphamide as described previously (30), inoculated the BALB/cJ mice with 7,000 PFU in 30 μl at day 4 after the initiation of drug treatment, and treated the mice with M16 at 2 dpi. Different groups treated with M16 were depleted of T cells (at days 11 and 14), NK cells (at days 4, 6, 10, 12, and 14), or both. Compared to the outcome in M16-treated mice, depletion of NK cells in M16-treated mice delayed viral clearance by approximately 7 days (Fig. 9A to C) and greatly increased the rate of morbidity (Fig. 9D). T-cell depletion combined with M16 treatment delayed the clearance by approximately 12 days compared to that seen in the M16-treated mice and increased the levels of lung infection and weight loss. M16 administration with T-cell depletion, however, greatly reduced the levels of nasal, tracheal, and lung infection relative to that observed in the T-cell-depleted group and resulted in 100% survival instead of 100% mortality (Fig. 9A to E). M16 administration with depletion of both NK and T cells had the same effect on bioluminescence as depletion of T cells alone, but it enhanced the rate of morbidity and caused 40% mortality (Fig. 9A to E). This experiment shows that M16 antibody treatment requires both NK and T lymphocytes for efficiency and proves that NK cells have critical immunoregulatory functions, even in mice treated with M16. Overall, the adoptive transfer of NK cells combined with M16 antibody treatment did not improve the therapeutic potential of M16 alone; however, NK cells were shown to be important for the maximal activity of M16.

FIG 9.

NK and T cells are required for an optimal antibody therapeutic effect. BALB/c mice were treated with dexamethasone (Dexa) in 10 daily doses of 10 mg/kg/day and cyclophosphamide (Cy) in 2 doses of 150 mg/kg at days 0 and 6. Mice were infected with 7,000 PFU of SeV in 30 μl PBS 4 days after the start of drug administration. The mice were treated with an anti-SeV monoclonal antibody (M16) 2 days after infection (day 6 after the initiation of drug treatment), administered as a single dose of 230 μg/mouse. Anti-CD3 antibody was administered as a dose of 20 μg/mouse at days 11 and 14, and anti-NK antibody (anti-ASGM1) was administered as a dose of 25 μl/mouse at days 4, 6, 10, 12, and 14 after the initiation of drug treatment. (A to C) Bioluminescence measured after i.p. injection of 150 mg/kg d-luciferin and imaging of the nasal cavity (A), trachea (B), and lungs (C) with a Xenogen system; (D) weight loss; (E) survival; (F to I) total counts of peripheral CD4⁺ cells (F), CD8⁺ cells (G), B cells (H), and NK cells (I) isolated at days 15 and 22 after drug initiation. The error bars represent the standard deviations, and there were 5 mice per group. The data on the x axes in panels A to E are the number of days after the transplant. P values were determined by Student's t test. *, P < 0.05; ***, P < 0.001.

Immunotherapy with T cells.

For the shared MHC haplotypes needed for T-cell immunotherapy, we used a haploidentical allogeneic HCT model with BALB/cJ mouse recipients and CB6F1/J mouse donors (CB6F1/J mice are the offspring of female BALB/cJ mice and male C57BL/6J mice). We infected the mice with 700 or 7,000 PFU of SeV in 30 μl, performed HCT, and on day 10 performed the adoptive transfer of T lymphocytes from the BALF of SeV-infected CB6F1/J mice. Adoptive T-cell transfer yielded significant increases (P < 0.05) in the number and percentage of SeV-specific CD8⁺ T lymphocytes, while it caused reductions in the proportion of CD4⁺ T cells, B cells, and NK cells (Fig. 10H, I, L, and M). Donor T-lymphocyte chimerism was also significantly increased (P < 0.05) (Fig. 10J and N). Adoptive T-cell transfer cleared the infection throughout the respiratory tract 5 days earlier in the mice receiving the T-cell transfer than in mice in the control group, regardless of whether the mice were infected with 700 or 7,000 PFU (Fig. 10). We next compared adoptive transfer using CD4⁺, CD8⁺, or total T lymphocytes. The transfer of either CD8⁺ or total T lymphocytes led to increases in the peripheral CD8⁺ T-lymphocyte populations (Fig. 11F) and a similar acceleration of clearance and reductions in the rates of morbidity and mortality (Fig. 11A to D). Overall, despite a relatively late treatment on day 10, T-cell adoptive transfer showed a remarkable therapeutic potential, largely due to CD8⁺ T cells.

FIG 10.

T-cell adoptive transfer accelerates SeV clearance after the transplant. BALB/cJ mice were irradiated, infected with 700 (A to C and G to J) or 7,000 (D to F and K to N) PFU of SeV in 30 μl PBS, and for transplantation given a T-cell-depleted bone marrow graft derived from CB6F1/J mice. T cells derived from the BALF of CB6F1/J mice that had been infected with SeV for 10 days were adoptively transferred 10 days after the transplant at a dose of 1 × 10⁵ cells/mouse delivered intravenously. (A to F) Bioluminescence measured after i.p. injection of 150 mg/kg d-luciferin and imaging of the nasal cavity (A and D), trachea (B and E), and lungs (C and F) with a Xenogen system; (G and K) peripheral lymphocyte counts at day 21 after the transplant; (H and L) peripheral SeV-specific CD8⁺ T cells at day 21 after the transplant; (I and M) flow cytometry of peripheral blood at day 21 after the transplant; (J and N) percent chimerism at day 21 after the transplant. The error bars represent the standard deviations, and there were 5 mice per group. P values were determined by Student's t test. *, P < 0.05; ***, P < 0.001.

FIG 11.

SeV-specific CD8⁺ T-cell adoptive transfer accelerates SeV clearance after HCT, reducing morbidity and mortality. BALB/cJ mice were irradiated, infected with 7,000 PFU of SeV in 30 μl, and for transplantation given a T-cell-depleted bone marrow graft derived from CB6F1/J mice. Total (orange diamonds), CD4⁺ (brown circles), or CD8⁺ (green triangles) T cells derived from the BALF of CB6F1/J mice that had been infected with SeV for 10 days were adoptively transferred intravenously 10 days after the transplant. Total T lymphocytes derived from the BALF of CB6F1/J mice that had been infected with influenza A/Puerto Rico/8/34 (H1N1) virus for 10 days (PR8⁺, blue triangles) were adoptively transferred as a control. (A to C) Bioluminescence in the nasopharynx (A), trachea (B), and lungs (C); (D and E) clinical effects of treatment in terms of percent weight change (D) and survival (E); (F) total peripheral CD8⁺ T cells at day 20 after the transplant as measured by flow cytometry. The error bars represent the standard deviations, and there were 5 mice per group. P values were determined by Student's t test. *, P < 0.05; ***, P < 0.001.

DISCUSSION

We developed a system to study the kinetics of respiratory virus spread, immunologic reconstitution, and infection clearance in mice undergoing HCT. We then investigated the clearance mechanisms in the transplant recipients and evaluated 4 candidate immunotherapies: G-CSF treatment, monoclonal antibody treatment, adoptive NK-cell transfer, and adoptive T-cell transfer. For these studies, we used a nonattenuated Sendai reporter virus so that we could study viral immunology using a matched host-pathogen pair (28–30, 32). SeV infection in mice has many similarities to parainfluenza virus infection in humans (24), which is highly prevalent in HCT recipients (5). By adjusting the inoculation dose of SeV and its volume, we could modulate the extent of lung infection and disease severity, as observed by the amount of weight loss, the rate of mortality, and the levels of T cells in the lungs and peripheral blood. This provided both survival and nonsurvival (challenge) models of infection. Increasing the donor-recipient mismatch increased the susceptibility to GVHD, thereby increasing viral pathogenesis. Noninvasive imaging of prolonged SeV infection in the transplant recipients showed that T cells play a central role in viral clearance. Our system should be extendable to other respiratory viruses, such as influenza virus and RSV, which are also prevalent in HCT recipients and for which reporter viruses have been developed (42–44). Clinically, our findings support the development of T-cell and antibody therapies for HPIV infection in HCT recipients, and they are significant because preclinical studies and demonstrated therapies are sorely lacking for this vulnerable population. These therapies may also be efficacious for the treatment of PIV, RSV, and influenza virus infections in other settings of immunocompromise, such as after chemotherapy, after solid-organ transplantation, and in neonates and elderly individuals.

In addition to testing candidate therapeutics, our study is the first to analyze the kinetics of respiratory virus infection and clearance and their relation to host responses in a mouse model of allogeneic HCT. During the first 5 days after inoculation, the initial increase in the level of SeV infection was similar to that seen in immunocompetent mice and in mice immunosuppressed by HCT in this study or by cyclophosphamide treatment in a previous study (30). The only notable exception to this was a small but significant increase in the level of nasal infection in the transplant recipients starting on day 3. An increased susceptibility to nasal infection in transplant recipients could be due to more severe lymphopenia. In immunocompetent mice, SeV clearance begins after 5 days of infection (30). This coincides with the start of T-cell infiltration and local and systemic antibody responses (28, 45). After HCT, the start of SeV clearance was delayed to approximately 21 days after the transplant, regardless of whether the lung infection was mild or severe or whether mice were fully MHC mismatched or partially MHC matched. SeV clearance after the transplant coincided with the recovery of lymphocytes, as observed in both immunocompetent and cyclophosphamide-treated mice (30). In the transplant recipients, the kinetics of the antibody responses lagged T-lymphocyte infiltration and remained relatively low until the infection was cleared. Impaired CD8⁺ T-cell function has recently been linked to prolonged and more severe SeV infection after allogeneic BMT (27). We demonstrated a central role for T lymphocytes in SeV clearance in the HCT recipients by antibody depletion experiments, in which a reduction in T cells but not a reduction in B cells prevented clearance. As there are no licensed vaccines, prophylactic antibodies, or antivirals against the HPIVs, the identification of T lymphocytes as the central driver of clearance in transplant recipients may help guide future supportive efforts aimed at reducing immune suppression after HCT (9).

The demonstrated efficacy of T-cell immunotherapy against SeV infection in the transplant recipients, with no increase in the rate of morbidity being associated with T-cell transfer, supports the clinical use of adoptive T-cell therapy for HPIV infection. On the basis of the findings of our preclinical studies, the expected advantages of T-cell immunotherapy over alternative approaches, such as treatment with monoclonal antibodies, would be the ability to control infection throughout the respiratory tract, as opposed to only the lungs, and a broad therapeutic window, which we have shown to extend to 14 days after infection. Delayed T-cell engraftment after HCT is a major risk factor for enhanced disease due to respiratory virus infections (46–49). Although we are unaware of any clinical studies on T-cell immunotherapy in HPIV-infected transplant recipients, such approaches have appeared promising for treating infections caused by several other viruses, including EBV, CMV, and adenovirus (50–52). These studies derived virus-specific T cells primarily from the donor and expanded them in vitro. Other promising approaches currently in clinical trials include the development of genetically modified T cells, multivirus-specific T cells, and T-cell banks with third-party T cells (53).

The prophylactic use of palivizumab, an RSV-specific monoclonal antibody against the F protein, in high-risk and immunocompromised patients reduces rates of hospitalization (54, 55). Comparable mouse monoclonal antibodies against the HPIV3 F and hemagglutinin-neuraminidase (HN) glycoproteins have been characterized in vitro (56–58), although it is unclear whether these antibodies have been humanized or have been advanced to clinical development. Our preclinical studies demonstrated that treatment with an anti-F monoclonal antibody halted the progression of SeV infection in the lungs and reduced signs of disease in mice that underwent HCT when it was administered as late as 10 dpi, providing preclinical support for the use of anti-HPIV antibodies as therapeutics. Early administration of antibody may have a role when HCT in a patient with HPIV infection cannot be delayed owing to the risk of progression of the underlying disease.

Increasing the neutrophil counts by administering G-CSF reduced the efficacy of the monoclonal antibody, perhaps as a result of the inflammatory potential of neutrophils. As G-CSF is routinely used in the transplant setting to hasten hematopoietic recovery (59, 60), combination therapy with G-CSF and an anti-HPIV monoclonal antibody might be investigated clinically for synergistic or antagonistic interactions.

Combining NK-cell adoptive transfer with a monoclonal antibody was neither advantageous nor deleterious compared to the effect of monoclonal antibody therapy alone, although the presence of NK cells was required for the maximal activity of M16 in our NK-cell depletion experiment. Ultimately, the best control of infection may be afforded by a combination of therapies that eliminate infected cells, such as T-cell adoptive transfer, and neutralize infectious virus, such as a monoclonal antibody, HN inhibitor (61–63), or sialic acid destroyer (64). Currently, there is no approved therapy for HPIV infection. Patients receive supportive care with oxygen therapy and intravenous hydration, and the rate of mortality is high.

NK cells are a first line of defense against viruses and tumors. As NK cells do not require previous exposure to antigens or MHC matching or cause GVHD, they could be used therapeutically in transplant recipients if they have intrinsic antiviral properties or can be stimulated to have such properties. NK-cell immunotherapy is currently used in the treatment protocols for retinoblastoma and in clinical trials for the treatment of acute myeloid leukemia, solid tumors, and lymphoma, among others (65). In most cases, killer cell immunoglobulin-like receptor (KIR) mismatch (36) is required for optimal tumor-killing activity. Previously, we showed that NK cells play a role in regulating SeV clearance and morbidity after immunosuppression with dexamethasone and cyclophosphamide (30). Here, we show the potential efficacy of NK-cell therapy when it is administered early after the transplant in the setting of a moderate viral load.

In summary, we developed a noninvasive system for imaging SeV infection in mouse transplant recipients and used it to reveal a central role for CD8⁺ T cells in viral clearance and to identify adoptive T-cell transfer and monoclonal antibodies as promising therapeutic candidates. The mouse model of respiratory infection in HCT recipients can be used to compare the therapeutic window and relative efficacies of candidate therapeutics against PIV. The model may also be extended to other RNA viruses that commonly infect this patient population, such as RSV and influenza A virus. Eligible patients may also possibly include recipients of high-dose chemotherapy and solid-organ transplants and those with primary immunodeficiency. The results of our study support the development of T-cell immunotherapeutics and virus-specific monoclonal antibodies against HPIV infection, for which there are currently no vaccines or virus-specific treatment options available in the clinic.

MATERIALS AND METHODS

Ethics statement.

All animal studies were conducted following the Guide for the Care and Use of Laboratory Animals (66) from the National Research Council of the National Academies of the United States. Experiments using mice were approved by the Animal Care and Use Committee of St. Jude Children's Research Hospital (protocol number 459). They were performed in compliance with relevant institutional policies, the Association for the Accreditation of Laboratory Animal Care guidelines, the National Institutes of Health regulations, and local, state, and federal laws. Animals were anesthetized with isoflurane and euthanized with CO₂.

Mice and hematopoietic cell transplants.

BALB/cJ recipient mice were maintained on water medicated with sulfamethoxazole and trimethoprim (280 mg sulfamethoxazole and 56 mg trimethoprim in a 350-ml water bottle), which were given 1 week before irradiation and up to 28 days thereafter. BALB/cJ mice received 2 doses of 500 cGy via a cesium-137 irradiator, with dose administration being separated by an interval of 3 to 4 h. After 24 h, mice received a transplant of 3 × 10⁶ to 5 × 10⁶ T-cell-depleted bone marrow cells from either C57BL/6J or CB6F1/J donor mice. To prepare the graft, the donors were euthanized by cervical dislocation while they were under anesthesia. Bone marrow was isolated from the tibias and femurs by flushing each bone twice with 1 ml of RPMI supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin, and 1% streptomycin, using 25-gauge needles. Cells were passed through 40-μm-pore-size filters, and T-cell depletion was performed using CD90.2 MicroBeads (Miltenyi Biotec) in accordance with the manufacturer's protocol. T-cell depletion was confirmed by flow cytometry.

Bioluminescence imaging of mice.

Female BALB/cJ (H-2^d) mice, C57BL/6J (H-2^b) mice, and CB6F1/J mice (CB6F1/J mice are the offspring of a cross between female BALB/c mice and male C57BL/6 mice) aged 6 to 8 weeks were purchased from The Jackson Laboratory (Bar Harbor, ME). Mice that underwent HCT were infected with the reporter Sendai virus rSeV-luciferase(M-F*) 6 h before they underwent a bone marrow transplant. For the infections, mice were anesthetized with isoflurane and inoculated intranasally with 5 or 30 μl of phosphate-buffered saline (PBS) or PBS containing 700 or 7,000 PFU of the virus. The animals were monitored daily for weight loss, morbidity, and mortality. Infected mice received an intraperitoneal (i.p.) injection of d-luciferin (PerkinElmer) at 150 mg/kg in 200 μl and were anesthetized with isoflurane for 5 min. Images were acquired with a Xenogen IVIS charge-coupled-device camera system (Caliper Life Sciences) and analyzed with Living Image (version 4.5) software (Caliper Life Sciences). To quantify the bioluminescence, regions of interest (ROIs) were defined manually, and the data were expressed as the total flux (number of photons per second).

Drugs, antibodies, and reagents.

Recombinant human G-CSF (Neupogen; Amgen Inc.) was administered by i.p. injection in 10 doses, each of which was 125 μg/kg in 200 μl, starting at day 5 after infection. The anti-mouse CD3e clone 145-2C11 (BioXcell) was administered in 2 i.p. doses of 10 μg in 100 μl PBS at days 13 and 16. The Armenian hamster IgG was used as an isotype control (BioXcell). Anti-asialo GM1 (anti-ASGM1) antibody (Wako Chemicals) was administered by i.p. injection to each mouse as 25 μl of the antibody in a total volume of 100 μl. Normal rabbit serum was used as a control (MP Biomedicals). The SeV monoclonal antibody M16, produced and purified by Envigo, was administered by i.p. injection at a dose of 230 μg per mouse. Mouse IgG (BioXcell) was used as an isotype control. Anti-mouse CD20 (Genentech) was administered by i.p. injection at 200 μg in a 100-μl volume at days 13 and 20, and mouse IgG2a (BioXcell) was used as an isotype control. To prepare the IL-15 and IL-15Rα complex, 20 μg IL-15 (Affymetrix) was mixed with 90 μg IL-15Rα (R&D systems) in 400 μl PBS at 37°C for 20 min. Each mouse received 1 μg in 100 μl PBS on the assigned days. Dexamethasone and cyclophosphamide were purchased from Sigma-Aldrich. Dexamethasone was administered by i.p. injection in 10 daily doses, each of which was 10 mg/kg/day in 200 μl. Cyclophosphamide was administered by i.p. injection in 2 doses of 150 mg/kg in 300 μl.

Blood and BALF collection and cell counts.

For blood collection, mice were anesthetized with isoflurane, and blood was collected in 10% EDTA (disodium) via the retro-orbital route. The cells were counted with an automatic cell counter. For bronchoalveolar lavage fluid (BALF) collection, mice were euthanized, and then their tracheas were intubated with 18-gauge catheters and washed 3 times with 0.5 ml of PBS containing 2 mM EDTA, giving a total lavage volume of 1.5 ml. The BALF was centrifuged, the supernatants were collected for enzyme-linked immunosorbent assay (ELISA), and the cellular pellet was resuspended in 50 to 100 μl of PBS. The cells were counted with an automatic cell counter.

Flow cytometry.

Blood or BALF cells, collected as described earlier, were blocked in mouse BD Fc Block (BD Biosciences) in staining medium (PBS plus 0.5% FBS) for 10 min at 4°C and then stained with B220-V500, CD4⁻ eF605, H2Kd-fluorescein isothiocyanate, H2Kb-phycoerythrin (PE), CD3⁻ Brilliant Ultraviolet 395 (BUV395), CD8⁻ Alexa Fluor 700, and CD49b⁻ allophycocyanin-Cy7 antibodies (BD Biosciences) for 30 min at 4°C. PE-conjugated Kb NP324-332 tetramer was used to detect SeV-specific CD8⁺ T cells, and staining was performed at room temperature for 1 h in combination with CD8-APC. Red blood cells were lysed using BD FACS lysing solution. Cellular populations were gated and quantified by using FlowJo (version 7.6.1) software.

NK-cell isolation and adoptive transfer.

Spleens were collected after the mice were euthanized by cervical dislocation while they were under anesthesia. The spleens were passed twice through 40-μm-pore-size filters. NK cells were isolated by using an NK-cell isolation kit II (Miltenyi Biotec) in accordance with the manufacturer's protocol. IL-15 at a concentration of 50 ng/ml was added to the isolation buffer. NK-cell purity was confirmed by staining with antibody against CD49b, and the enrichment was between 70% and 90%. Cells were adoptively transferred intravenously at 1 × 10⁷ cells per mouse. The IL-15–IL-15R complex was injected intraperitoneally 1 day before, 1 day after, and 3 days after NK-cell transfer.

T-lymphocyte isolation and adoptive transfer.

BALF was collected from CB6F1/J mice infected with SeV for 10 days as described above. Cell pellets were collected, and total, CD4⁺, and CD8⁺ T cells were isolated by using an Untouched mouse T-cell kit (Invitrogen), a CD8a T-cell isolation kit (Miltenyi Biotec), and a CD4 T-cell isolation kit (Miltenyi Biotec) in accordance with the manufacturers' instructions. Of the isolated T cells, 70% showed positive staining by the NP324-332 tetramer. Each mouse received 1 × 10⁵ cells injected intravenously.

Histopathology.

Nasal cavities, tracheas, and lungs were collected in 10% neutral buffered formalin. Paraffin embedding, sectioning, and staining of tissues were performed by the Veterinary Pathology Core Laboratory at St. Jude Children's Research Hospital.

ELISA.

For ELISA, 96-well plates were coated overnight with disrupted, purified SeV (10 μg/ml). The plates were blocked with PBS containing 1% bovine serum albumin and then incubated with 10-fold serially diluted serum samples or undiluted BALF supernatants for 1 h at 37°C. After incubation, the plates were washed, incubated with goat anti-mouse IgG or IgA conjugated to alkaline phosphatase (AP; Southern Biotechnology) for 1 h at 37°C, and then washed further. Next, p-nitrophenyl phosphate substrate was added to the wells (Sigma), and the plates were read at a wavelength of 405 nm.

Quantification and statistical analysis.

Statistical parameters, including the exact value of the number of animals, standard deviations, and significance, are reported in the figures and figure legends. Data were considered significant if P was <0.05 by 2-tailed Student's t test or 1-way analysis of variance (ANOVA) followed by Bonferroni posttests. Statistical analysis was performed using Microsoft Excel or GraphPad Prism software by an investigator not involved in performing the experiments.