Functional Antagonism of Chemokine Receptor CCR1 Reduces Mortality in Acute Pneumovirus Infection In Vivo

Abstract

We present an antiviral-immunomodulatory therapeutic strategy involving the chemokine receptor antagonist Met-RANTES, which yields significant survival in the setting of an otherwise fatal respiratory virus infection. In previous work, we demonstrated that infection with the natural rodent pathogen pneumonia virus of mice involves robust virus replication accompanied by cellular inflammation modulated by the CC chemokine macrophage inflammatory protein 1α (MIP-1α). We found that the antiviral agent ribavirin limited virus replication in vivo but had no impact on morbidity and mortality associated with this disease in the absence of immunomodulatory control. We show here that ribavirin reduces mortality, from 100% to 10 and 30%, respectively, in gene-deleted CCR1^−/− mice and in wild-type mice treated with the small-molecule chemokine receptor antagonist, Met-RANTES. As MIP-1α-mediated inflammation is a common response to several distantly related respiratory virus pathogens, specific antiviral therapy in conjunction with blockade of the MIP-1α/CCR1 inflammatory cascade may ultimately prove to be a useful, generalized approach to severe respiratory virus infection and its pathological sequelae in human subjects.

As of this writing, we are still without safe and effective therapeutic strategies for acute respiratory virus infections. Despite considerable efforts over the almost 50 years since its first description, supportive therapy alone remains the standard of care for the treatment of severe cases of respiratory syncytial virus (RSV), a disease with significant morbidity and mortality, particularly among infants born prematurely, and for which there is currently no vaccine (5, 8). Even when a vaccine is available for a respiratory pathogen, such as for epidemic influenza virus, similar problems of disease prevention and management exist despite seasonal reformulation of vaccine preparations (17, 25). While progress has been made toward the development of small-molecule agents with antiviral activity in vitro, the clinical impact of these therapies has been disappointing (1). At the source of the problem is the observation that, once established, respiratory virus disease results from two concurrent pathological components: ongoing virus replication and the resulting inflammatory response. Even when antivirals clearly inhibit virus replication, the biochemical and cellular inflammatory responses to the initial infection-related events continue despite diminished virus titer (4, 26). While acute inflammatory responses are generally beneficial in nature and have been shown to limit virus replication in situ (10), prolonged, uncontrolled inflammation has been recognized as a significant component contributing to the pathological sequelae of RSV and influenza virus, and most recently, to the morbidity and mortality of severe acute respiratory syndrome coronavirus infection (16, 19).

Chemokines and chemokine receptors orchestrate the recruitment of proinflammatory leukocytes, and as such represent interesting and important targets for the subversion of antiviral inflammatory responses and leukocyte-mediated tissue damage. Acute inflammatory responses to RSV, a pneumovirus of the family Paramyxoviridae, and Influenza virus, of the family Orthomyxoviridae, are mediated by the CC chemokine macrophage inflammatory protein-1α (MIP-1α/CCL3). In human studies of infants on ventilatory support as a result of severe RSV infection, MIP-1α was detected in bronchoalveolar lavage (BAL) fluid at levels correlating with leukocyte degranulation products (14) and with severity of disease (12). MIP-1α was likewise among the chemokines produced by human monocytes infected with influenza A virus (15), and MIP-1α^−/− mice demonstrated attenuated inflammatory responses to acute influenza infection (6). While the precise interplay of and/or hierarchy among chemokine cascades remains to be defined, the data suggest that specific blockade of MIP-1α signaling might reduce the impact of prolonged inflammation observed in response to respiratory virus infections in general. When administered in conjunction with appropriate antiviral therapy, specific blockade of MIP-1α signaling might ultimately result in an improved clinical outcome.

We have developed and characterized a mouse model of severe, acute respiratory infection using the highly virulent natural rodent pathogen Pneumonia virus of mice (PVM), family Paramyxoviridae, subfamily Pneumovirinae, strain J3666. Mice that are inoculated with fewer than 100 PFU develop acute lower airway disease characterized by progression to pneumonia, respiratory failure, and death (9, 10). We have previously demonstrated that, analogous to influenza and RSV infection, MIP-1α is central to this virus-induced inflammatory process and that interruption of MIP-1α signaling via the receptor CCR1 results in marked reduction in pulmonary inflammation (10). Furthermore, we have shown that the inflammatory events associated with pneumovirus infection are not inextricably linked to ongoing viral replication, as once an infection as been initiated, subsequent treatment with ribavirin does not put a stop to the associated inflammatory events contributing to the infection-associated morbidity and mortality (4).

The experiments described here demonstrate that the biochemical chemokine receptor antagonist Met-RANTES achieves functional antagonism of MIP-1α/CCR1 signaling and results in a pronounced attenuation of cellular inflammation associated with pneumovirus infection in vivo. When combined with the potent antiviral compound ribavirin, Met-RANTES provides dramatic, dose-dependent reduction in morbidity and mortality in response to severe respiratory virus infection.

MATERIALS AND METHODS

Mouse and virus stocks.

Wild-type C57BL/6 mice were obtained from Taconic Laboratories, Germantown, N.Y. Gene-deleted CCR1^−/− mice on the C57BL/6 background were developed and bred on-site at the National Institutes of Health (11). Mouse-passaged stocks of PVM (strain J3666, 1.5 × 10⁶ PFU/ml) were obtained from clarified mouse lung homogenates as described previously (9) and stored in liquid nitrogen. Virus stocks were defrosted and diluted in phosphate-buffered saline (PBS) to 120 PFU/100 μl (>1,000-fold) immediately prior to intranasal inoculation. Inoculation of mice with UV-inactivated virions at this dilution does not result in leukocyte recruitment nor in production of MIP-1α (data not shown). All procedures were reviewed and approved by the National Institutes of Health Animal Welfare Review Board (protocol LHD8E) and/or the Committee on the Humane Use of Animals, SUNY Upstate Medical University (CHUA #634).

Establishing PVM infections in mice and treatment interventions.

Six to 8-week-old mice were used in all experiments. Mice subjected to brief inhalation anesthesia with 20% isoflurane in mineral oil were inoculated intranasally with 60 PFU of mouse-passaged PVM strain J3666 in a 50-μl volume of PBS at day 0. All treatment interventions were initiated on day 3. Clinical scoring of infected mice was performed in a blinded fashion via a scoring method initially devised by Easton and colleagues (7) with modifications (4) as follows: 1, healthy; 2, ruffled fur at neck; 3, piloerection and difficulty breathing, less alert; 4, lethargic with labored breathing; 5, premorbid, with emaciation and cyanosis; 6, death. Macroscopic lung scores also obtained in blinded fashion were determined as follows: 1, normal; 2, blushed or gray without focal lesions; 3, unifocal lesion on either lung; 4, multifocal lesions on one lung; 5, multifocal lesions on both lungs; and 6, grossly hepatized, with or without visual honeycombing.

Treatment interventions included (i) intraperitoneal ribavirin (37.5 mg/kg of body weight/dose), two doses per day, 1.0 to 1.5 ml each, separated by at least 6 h (Sigma Aldrich, St. Louis, Mo.), or an equal volume of diluent control (PBS) and/or (ii) Met-RANTES (1, 10, or 100 μg intraperitoneallyper mouse per day in a 1-ml volume; generously provided by Serono Pharmaceutical, Geneva Switzerland), or an equal volume of diluent control (PBS).

BAL and differential cell counts.

At the time points indicated, mice were sacrificed by cervical dislocation (a minimum of three mice per condition per time point) and BAL fluids were harvested by trans-tracheal instillation and removal of prechilled phosphate-buffered saline with 0.25% bovine serum albumin (two 0.80-ml instillations with recovery of 1.2 to 1.5 ml per mouse). Total and differential leukocyte counts were obtained by light microscopic quantitative analysis of methanol-fixed cytospin preparations stained with DiffQuik (Fisher Scientific, Pittsburgh, Pa.).

Microscopic and biochemical analysis and virus plaque assays of lung homogenates.

Mice were sacrificed as described above (minimum of three mice per condition per time point). Lungs to be sectioned for microscopic analysis were inflated trans-tracheally with 0.3 ml 10% formalin before removal from the body cavity. Inflated lungs were fixed in 10% formalin before sectioning, slide preparation, and staining with hematoxylin and eosin (American Histolabs, Gaithersburg, Md.). Lungs to be processed for enzyme-linked immunosorbent assay and viral titer determination were removed and transferred into 2 ml of prechilled Iscove's modified Dulbecco's medium. Lung tissue suspensions were subjected to blade homogenization (Tissumizer; Tekmar, Cincinnati, Ohio) and cellular debris was removed by low-speed centrifugation (500 × g at 4°C), yielding homogenates at 7 to 12 mg protein/ml. Clarified supernatants were flash frozen in a dry ice-ethanol slurry and stored at −80°C prior to analysis. Assays for mouse MIP-1α were performed as per the manufacturer's instructions (R&D Systems, Minneapolis, Minn.) and results were corrected for total protein as determined by the Bradford colorimetric assay using bovine serum albumin standards. Viral recovery was determined by standard plaque assay on the BS-C-1 epithelial cell line (American Type Culture Collection, Manassas Va.).

Statistical analysis.

Data points represent the mean ± SE (standard error) of samples from three or more trials. Fisher's exact test was employed for categorical (clinical) data. Unpaired t tests were used to compare continuous data as per the algorithms of the Microsoft Excel data analysis program. Kaplan Meier Analyses were performed using Statistica Software (StatSoft, Tulsa, Okla.).

RESULTS

Therapeutic interventions and virus titers.

Virus titers were determined in lung tissue homogenates from mice receiving Met-RANTES (1, 10 or 100 μg/day), ribavirin (75 mg/kg/day in two divided doses), or both, beginning on day 3 postinoculation (Table 1). Virus titers increased over time throughout, reaching 1 × 10⁸ to 2 × 10⁸ PFU/g lung tissue by day 7 among mice that did not receive ribavirin. Ribavirin therapy resulted in a marked inhibition of virus replication, with >800- to >1,500-fold reductions in virus titer observed on days 5 and 7 postinoculation, respectively (P < 0.001). Met-RANTES had no impact on virus replication.

TABLE 1.

Ribavirin-mediated inhibition of virus replication in vivo^a

Treatment	PVM titer (PFU/g of lung tissue [106]) on day:
	0	3	5	7
None	ND	0.6 ± 0.02	34 ± 10	143 ± 60
Met-RANTES (1 μg/day)	ND	0.8 ± 0.01	40 ± 9	180 ± 38
Met-RANTES (10 μg/day)	ND	0.8 ± 0.02	44 ± 7	166 ± 45
Met-RANTES (100 μg/day)	ND	0.5 ± 0.01	38 ± 11	173 ± 30
Ribavirin	ND	0.7 ± 0.01	0.04 ± 0.007* [850]	0.09 ± 0.004* [1,590]
Met-RANTES (1 μg/day) + ribavirin	ND	0.8 ± 0.03	0.02 ± 0.004* [2,000]	0.07 ± 0.003* [2,570]
Met-RANTES (10 μg/day) + ribavirin	ND	0.6 ± 0.02	0.03 ± 0.01 [1,470]	0.09 ± 0.01* [1,844]
Met-RANTES (100 μg/day) + ribavirin	ND	0.7 ± 0.03	0.04 ± 0.01 [950]	0.08 ± 0.02* [2,160]

^aNumber of PFU was determined by standard assay from lung homogenates from PVM-infected wild-type mice treated with Met-RANTES and/or ribavirin (75 mg/kg/day) or diluent controls beginning on day 3; n = 6 mice per data point. Statistical significance, *P < 0.001 compared to the same concentration of Met-RANTES, no ribavirin control at same time point. Fold reduction is shown in brackets; ND, none detected.

Therapeutic interventions and production of MIP-1α and leukocyte recruitment to lung tissue of virus-infected mice.

MIP-1α was detected in lung tissue in virus-infected mice beginning on day 5 postinoculation (Table 2). Met-RANTES-mediated receptor blockade had no effect on MIP-1α levels, nor did we observe altered production of MIP-1α in mice with complete blockade of CCR1 (i.e., in CCR1^−/− mice). Interestingly, while administration of ribavirin resulted in dramatic reductions in virus titer (>800- to nearly 3,000-fold [Table 1]), its impact on MIP-1α production was minimal, with only two- to threefold reductions in MIP-1α levels detected at these time points. As such, it is not surprising that ribavirin treatment alone had no impact on leukocyte recruitment in response to virus infection (Table 3). Consistent with our earlier studies (4) these findings indicate that ongoing virus replication and the ensuing inflammatory response are not tightly linked to one another during the course of natural respiratory virus infection in vivo.

TABLE 2.

MIP-1α (pg/ml/mg protein) detected in lung homogenates of virus-infected mice^a

Treatment	3	5	7
+pbs	ND	98 ± 26	407 ± 87
+Met-RANTES (1 μg/day)	ND	85 ± 21	377 ± 74
+Met-RANTES (10 μg/day)	ND	94 ± 12	388 ± 58
+Met-RANTES (100 μg/day)	ND	82 ± 35	401 ± 80
+ribavirin	ND	29 ± 8* [3.4]	178 ± 34* [2.3]
+Met-RANTES (1 μg/day) + ribavirin	ND	36 ± 6* [2.4]	157 ± 12* [2.4]
+Met-RANTES (10 μg/day) + ribavirin	ND	41 ± 18* [2.3]	168 ± 37* [2.3]
+Met-RANTES (100 μg/day) + ribavirin	ND	23 ± 10* [3.6]	182 ± 70* [2.2]
CCR1−/− mice	ND	76 ± 14	371 ± 106

^aLung homogenates were prepared from PVM-infected wild type mice treated with ribavirin (75 mg/kg/day) and/or Met-RANTES at the concentrations indicated or CCR1 −/− mice (untreated) on days postinoculation as indicated. All treatments were initiated on day 3 postinoculation. Data are presented as mean ± SE; ND, none detected; *P < 0.05 compared to +pbs controls, n = 3 mice per data point. Fold reduction (vs. same Met-RANTES, no ribavirin control) is indicated in brackets.

TABLE 3.

Leukocytes in bronchoalveolar lavage fluid of virus-infected mice^a

Day	Leukocytes (102)/ml of BAL fluid ± SE
	No Met-RANTES		Met-RANTES (10 μg/day)		Met-RANTES (100 μg/day)		CCR1−/− mice
	+pbs	+ribavirin	+pbs	+ribavirin	+pbs	+ribavirin
3	3.3 ± 0.3	2.3 ± 0.6	1.8 ± 0.3	2.0 ± 0.5	0.8 ± 0.01	0.6 ± 0.01	0.3 ± 0.04
5	4,960 ± 270	3,770 ± 300	535 ± 48* [9.3]	486 ± 55* [7.7]	87 ± 9** [57]	99 ± 21** [38]	0.5 ± 0.1 [9,900]**
7	5,970 ± 480	5,080 ± 410	1,040 ± 75* [5.7]	1,220 ± 100* [4.2]	210 ± 23** [28]	265 ± 48** [19]	1.4 ± 0.3 [4,300]**
10	3,560 ± 250	4,670 ± 370	810 ± 44* [4.3]	990 ± 30* [4.8]	185 ± 39** [19]	175 ± 30** [27]	1.0 ± 0.5 [3,600]**

^aBronchoalveolar lavage (BAL) fluids were harvested from PVM-infected wild-type mice treated with ribavirin (75 mg/kg/day) or control (pbs) and/or Met-RANTES or CCR1 −/− mice (untreated) on days postinoculation as indicated. All treatments were initiated on day 3 postinoculation. Data are presented as mean ± SE; *P < 0.05 compared to +pbs or +ribavirin alone controls, respectively; **P < 0.05 compared to +pbs +Met-RANTES (10 μg/day) or +ribavirin +Met-RANTES +(10 μg/day) groups, respectively, n = 3 mice per data point. Fold reduction (vs. no Met-RANTES at same time point) is shown in brackets.

Met-RANTES, however, had a significant, dose-dependent effect on leukocyte recruitment, with Met-RANTES at 10 and 100 μg/day yielding 4- to 8-fold and 20- to 40-fold reductions in leukocyte recruitment to lungs of virus-infected mice, respectively (Table 3). Interestingly, complete signaling blockade via deletion of the gene encoding of CCR1 (CCR1^−/−) resulted in even more substantial reductions in leukocyte recruitment (4,000- to 10,000-fold).

Macroscopic lung pathology in virus-infected mice.

Gross pathological specimens were observed and scored on a tiered six-point system described in Materials and Methods and in the legend to Fig. 1. The images depict specific examples, scored as 1 (normal, Fig. 1A), 3 (single hemorrhagic lesion, Fig. 1B) and 5 (multiple hemorrhagic lesions, Fig. 1C). The macroscopic lung scores for each treatment group are shown in Fig. 1D, establishing that Met-RANTES (both the 10 or 100 μg/day doses) together with ribavirin resulted in diminished lung scores at day 5 and thereafter compared to PBS control, to Met-RANTES alone (100 μg/day), or to ribavirin alone (75 mg/kg/day; P < 0.05 for each comparison). A difference was also noted between the dual therapy groups receiving 10 and 100 μg of Met-RANTES and ribavirin per day at the day 10 time point only (P < 0.05).

FIG. 1.

Macroscopic pathology scored in lungs of PVM-infected, ribavirin and/or Met-RANTES treated wild-type mice. Pathology scores were tabulated as follows. Macroscopic lung scores were determined as follows: 1, normal; 2, blushed or gray without focal lesions; 3, unifocal lesion on either lung; 4, multifocal lesions on one lung; 5, multifocal lesions on both lungs; and 6, grossly hepatized, with or without visual honeycombing. Examples of gross pathology and scoring include normal, score 1 (A); single focal lesion on right lower lobe, score 3 (B); and multifocal lesions on both lungs, score 5 (C). (D) Data are presented as means ± standard errors (error bars); *, P < 0.05 compared to treatment with PBS, Met-RANTES alone, or ribavirin alone; **, P < 0.05 compared to the group treated with ribavirin plus Met-RANTES (10 μg).

Morbidity scoring of virus-infected mice.

Clinical morbidity was scored on a tiered six-point system as described previously (4) and in the legend to Fig. 2. Morbidity scores tabulated for the mice in the combined Met-RANTES (10 or 100 μg/day) and ribavirin therapy groups were significantly lower than for the control, ribavirin only, or Met-RANTES only groups on day 8 postinoculation and thereafter (P < 0.001). In a comparison of the combined therapy groups to one another, significant differences were observed on and after day 8 postinoculation (P < 0.01).

FIG. 2.

Morbidity scores of PVM-infected, ribavirin- and/or Met-RANTES-treated wild-type mice. Morbidity scores were tabulated as follows: 1, healthy; 2, ruffled fur at neck; 3, piloerection and difficulty breathing, less alert; 4, lethargic with labored breathing; 5, premorbid, with emaciation and cyanosis; 6, death. Data are presented as means ± standard errors (error bars); *, P < 0.01 compared to treatment with Met-RANTES alone, ribavirin alone, or PBS; **, P < 0.01 compared to the group treated with ribavirin plus Met-RANTES (10 μg).

Survival analysis.

As shown in Fig. 3, long-term survival (>14 days) for mice that received PBS (control) or single therapy with either Met-RANTES (100 μg/day) or ribavirin was ≤10%. Forty-five (45%) percent of the mice (9 of 20) receiving ribavirin and low-dose Met-RANTES (10 μg/day) survived (P < 0.01 versus control and single-therapy groups), while 70% (14 of 20) of the mice receiving ribavirin and high-dose Met-RANTES (100 μg/day) survived (P < 0.01 when compared to control and single-therapy groups, P < 0.05 when compared to the ribavirin and low dose Met-RANTES group). Survival of CCR1^−/− mice not receiving Met-RANTES is slightly delayed (P < 0.05) but similar to the analogous wild-type curve (10). Ninety percent (18 of 20) of the CCR1^−/− mice receiving ribavirin survived (P < 0.01 versus all other groups), indicating that improved survival correlates directly with the extent of anti-inflammatory blockade. All mice surviving through day 14 were carried through day 28, confirming long-term-survivor status.

FIG. 3.

Survival of PVM-infected, ribavirin- and/or Met-RANTES-treated wild-type mice. Significantly improved survival was seen in the groups treated with ribavirin and Met-RANTES (*, P < 0.05) compared individually to the groups treated with Met-RANTES alone, ribavirin alone, or PBS groups. The combination of ribavirin with the higher Met-RANTES dose (100 μg/day) offered additional survival benefits over those offered by the ribavirin-Met-RANTES (10 μg/day) group (**, P < 0.05), approaching that observed for CCR1^−/− mice treated with ribavirin, representing theoretical complete receptor blockade.

DISCUSSION

In this work, we have demonstrated the efficacy of functional antagonism of CCR1 as part of a therapeutic approach to severe respiratory virus infection. Specifically, we have shown that mice respond to the specific immunomodulatory blockade provided by the receptor antagonist, Met-RANTES with a blunted cellular inflammatory response to pneumovirus infection. When administered in conjunction with ribavirin, the combined antiviral-immunomodulatory approach yields significant reductions in morbidity and mortality in response to this otherwise fatal respiratory virus infection.

Chemokines orchestrate inflammation and, as such, represent logical targets for novel and specific anti-inflammatory therapies. As noted earlier, the chemokine MIP-1α modulates lung inflammation associated with respiratory virus infections, including lesions induced by influenza A virus and RSV (6, 12, 14, 15). The observation that these two distantly related respiratory viruses share a crucial proinflammatory signaling pathway supports the notion of MIP-1α/CCR1 as a more broadly relevant pathway mediating virus-induced inflammation, a finding that may have implications for other virus-associated severe, acute respiratory syndromes (13).

In this work, we have focused on a clinically relevant strategy for treating respiratory virus infections in vivo. The respiratory virus selected undergoes rapid and robust replication in lung tissue and results in easily measured and clinically relevant outcomes (4, 10). We have initiated therapies on day 3 postinoculation so as to permit the virus to replicate unabated for several days, again reflective of a clinically relevant situation. Similarly, we have focused on a small molecule biochemical antagonist as a therapeutic agent, specifically Met-RANTES, characterized as a functional antagonist of chemokine signaling at CCR1 and CCR5 (20, 21). In this work, we have focused on Met-RANTES and its interactions with CCR1. While there may be the potential for additive and/or synergistic contributions from interactions with CCR5, it is interesting that none of the trials with Met-RANTES provided a level of protection that was superior to that observed in response to complete interruption of CCR1 signaling (i.e., in CCR1 gene-deleted mice). Met-RANTES in other therapeutic settings serves to abrogate posttransplant small bowel inflammation (2, 23), to attenuate chronic allograft nephropathy (22), to blunt the inflammation-mediated lung injury in a mouse model of chemical pancreatitis (3), to protect against Clostridium difficile toxin-mediated inflammatory colitis (18), and to limit dendritic cell infiltration into the airways in response to heat-killed bacteria (24). This is the first report documenting the use of this antagonist to reduce the lethal inflammatory responses accompanying severe respiratory virus infection.

In our earlier work on this subject (10), we determined that virus replication was enhanced in both MIP-1α^−/− mice and CCR1^−/− mice infected with PVM, with the finding in MIP-1α^−/− mice replicated in a subsequent study (4). In this study, despite the presence of inflammatory blockade, we see no enhanced virus replication at any time point upon administration of the small-molecule antagonist Met-RANTES (Table 1). While this may be related to the fact that Met-RANTES administration begins on day 3, while the effects of gene deletion exist throughout, we are unable to explain this difference to our complete satisfaction at the present time.

In summary, we have demonstrated that antagonism of signaling via the chemokine receptor CCR1 is a potent strategy that attenuates leukocyte recruitment in response to severe respiratory virus infection. Maximum clinical impact—significant reductions in morbidity and mortality in response to an otherwise fatal respiratory virus infection—was achieved by combining the specific functional antagonism at CCR1 with an effective antiviral agent. As MIP-1α-modulated inflammation is a common response to severe respiratory virus infection, specific antiviral therapy in conjunction with blockade of the MIP-1α/CCR1 may be feasible as a more generalized approach to severe respiratory virus infection and treatment of its pathological sequelae.