Transmission of Guanarito and Pirital Viruses among Wild Rodents, Venezuela

Secretions and excretions from virus-infected cane mice and cotton rats might transmit disease to humans.

Hemorrágica Venezolana) number and then killed by exposure to a lethal dose of vaporized chloroform. The identifi cation number, date of capture, trap site and trap number, species identity, sex, total length (tip of nose to tip of tail), length of tail, and other information were recorded on a standardized form. A throat swab; samples of blood, lung, spleen, liver, and kidney; and a sample of urine were collected from each rodent (6). These samples were stored in cryo-vials in liquid nitrogen in the fi eld and then shipped on dry ice to the University of Texas Medical Branch (Galveston, TX, USA).

Virus Assay
The samples from the throat swabs; crude 10% wt/vol homogenates of the samples of lung, spleen, and kidney in 0.01 mol/L phosphate-buffered saline; and samples of urine were assayed for arenavirus by cultivation in monolayers of Vero E6 cells (5). Cells harvested from the monolayers on day 13 or 14 postinoculation were tested for arenaviral antigen by using an indirect fl uorescent antibody test (IFAT) in which the primary antibody was a mixture of a hyperimmune mouse ascitic fl uid (HMAF) raised against the GTOV prototype strain INH-95551 (7) and an HMAF raised against the PIRV prototype strain VAV-488 (2).

Serologic Characterization of Viruses
Strains of GTOV were distinguished from strains of PIRV by an ELISA (5). The test antigens were detergent lysates of infected Vero E6 cells. Serial 2-fold dilutions (from 1:800 through 1:204,800 vol/vol) of an anti-GTOV HMAF and anti-PIRV HMAF were tested against each antigen. Antibody (IgG) bound to antigen was detected by using a goat antimouse (Mus musculus) IgG peroxidase conjugate in conjunction with the ABTS Microwell Peroxidase Substrate System (Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA). The reactivity of an HMAF against an antigen was the sum of the optical densities of the 8 reactions in the series of 4-fold dilutions of the HMAF tested against the antigen. The identity of an isolate was determined by direct comparison of the reactivity of the anti-GTOV HMAF versus the reactivity of the anti-PIRV HMAF against the test antigen.

Genetic Characterization of Viruses
The sequences of a 616-619-nt fragment of the nucleocapsid (N) protein genes of the arenaviruses isolated from the spleens of 21 rodents in this study (Table 1) were determined to assess the accuracy of the interpretation of the ELISA data. Total RNA was isolated from monolayers of infected Vero E6 cells by using TRIzol Reagent (Invitrogen Life Technologies, Inc., Carlsbad, CA, USA) or Tri Reagent (Sigma Aldrich, St. Louis, MO, USA). First-strand cDNA was synthesized by using SuperScript II RNase H -Reverse Transcriptase (Invitrogen Life Technologies, Inc.) in conjunction with oligonucleotide 19C-cons (8). Amplicons were synthesized from fi rststrand cDNA by using the MasterTaq Kit (Eppendorf North America, Inc., Westbury, NY, USA) in conjunction with oligonucleotides that fl ank either a 619-nt fragment of the N protein gene of GTOV strain INH-95551 or the homologous region (a 616-nt fragment) of the N protein gene of PIRV strain VAV-488. Amplicons of the expected size were sequenced directly by using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Inc., Foster City, CA, USA).

Antibody Assay
Blood samples were rendered noninfectious by irradiation (5 × 10 6 rads, Co 60 source), diluted 1:20 vol/ vol in phosphate-buffered saline, and then tested for IgG against GTOV strain INH-95551 and PIRV strain VAV-488 by using an IFAT. The cell spots were either a mixture of Vero E6 cells infected with INH-95551 and uninfected Vero E6 cells or a mixture of Vero E6 cells infected with VAV-488 and uninfected Vero E6 cells. Antibody bound to antigen was revealed by using a fl uorescein isothiocyanateconjugated goat antibody raised against mouse (M. musculus) IgG (Kirkegaard and Perry). End-point titers against INH-95551 and VAV-488 were measured in the positive samples by using serial 2-fold dilutions beginning at 1:20 and ending at 1:640 vol/vol.

Data Analysis
The male short-tailed cane mice, female short-tailed cane mice, male Alston's cotton rats, and female Alston's cotton rats were assigned on the basis of their nose-torump lengths (measured in mm) to 4 size categories. In each instance, the upper boundary of class I was the mean length -1 SD, the upper boundary of class II was the mean length, the upper boundary of class III was the mean length + 1 SD, and the upper boundary of class IV was the longest nose-to-rump length ( Table 2). Animals that were culture-positive or antibody-positive were treated as infected. The acceptable type I error in all statistical tests was α = 0.05.
Antibody titers <20 were considered 10 in comparisons of antibody titers to GTOV strain INH-95551 and PIRV strain VAV-488 in individual blood samples. The apparent homologous virus in an antibody-positive sample was the virus associated with the highest titer if the absolute value of the difference between the titers to GTOV and PIRV was >4-fold. ‡Pirital virus (PIRV) was isolated from all of the infected cotton rats, and antibody against PIRV was found in 1 of the culture-positive cotton rats. The table does not include FHV-4149, the only cotton rat infected with GTOV.

Results
A total of 128 rodents were captured on Hato Maporal in February 1997 in 1,000 trap-nights, with an overall trap success rate of 12.8% (Table 3). Most (121 [94.5%]) of the 128 rodents were short-tailed cane mice or Alston's cotton rats.
Fifty-seven (89.1%) of the 64 short-tailed cane mice and 55 (96.5%) of the 57 Alston's cotton rats were captured in 91 (37.9%) of the 240 traps set on sites A and B. Six Alston's cotton rats from site A were found in traps adjacent to traps in which cane mice were captured, 17 Alston's cotton rats from site B were found in traps adjacent to traps in which cane mice were captured, and a cotton rat and cane mouse were captured on different nights in each of 3 traps on site A and 6 traps on site B. Collectively, these observations suggest that the shorttailed cane mice captured on sites A and B lived in close proximity to Alston's cotton rats and vice versa.
Arenavirus was isolated from the throat swabs; samples of lung, spleen, or kidney; and samples of urine from 29 (45.3%) of the 64 short-tailed cane mice, 37 (64.9%) of the 57 Alston's cotton rats, and none of the 7 pygmy rice rats ( Table 4). The analyses of the ELISA data indicated that the arenaviruses isolated from Alston's cotton rat FHV-4149 and the short-tailed cane mice are strains of GTOV and that the arenaviruses isolated from the Alston's cotton rats other than FHV-4149 are strains of PIRV. The results of the neighbor-joining analysis of N protein gene sequence data (Figure) were 100% concordant with the serologic identities of the 21 viruses selected for genetic characterization, GTOV strain AV 97021119, and PIRV strain AV 97021016.
Nucleotide sequence nonidentity between the GTOV strains from 1994 (i.e., VAV-623 and VAV-952) was 6.0%, nucleotide sequence nonidentities among AV 97021119 and the 12 other GTOV strains from 1997 ranged from 0 to 7.3%, and nucleotide sequence nonidentities between the GTOV strains from 1994 and the 13 GTOV strains from 1997 ranged from 0.3% to 6.6%. Similarly, nucleotide sequence nonidentity between the PIRV strains from 1994 (i.e., VAV-628 and VAV-956) was 5.2%, nucleotide sequence nonidentities among AV 97021016 and the 9 other PIRV strains from 1997 ranged from 0 to 6.3%, and nucleotide sequence nonidentities between the PIRV strains from 1994 and the 10 PIRV strains from 1997 ranged from 1.0% to 6.3%. Antibody (IgG) against GTOV or PIRV was found in 11 (37.9%) of the 29 culture-positive cane mice, 2 (5.7%) of the 35 culture-negative cane mice, 1 (2.7%) of the 37 culture-positive cotton rats, none of the 20 culture-negative cotton rats, and none of the 7 pygmy rice rats ( Table 4). The only antibody-positive cotton rat (FHV-4124) was mature (size class III) and antibody positive to GTOV and PIRV. None of the 7 cane mice in size class I was antibody positive to GTOV or PIRV.
The end-point antibody titers to GTOV in the antibodypositive cane mice ranged from 40 to 160, none of the cane mice were antibody positive to PIRV, and the end-point antibody titers to GTOV and PIRV in the antibody-positive Alston's cotton rat were 40 and >640, respectively. Thus, GTOV was the apparent homologous virus in all of the antibody-positive cane mice, and PIRV was the apparent homologous virus in the antibody-positive cotton rat.
Eight (26.7%) of the 30 short-tailed cane mice in size classes I and II, 23 (67.6%) of the 34 cane mice in size classes III and IV, 11 (68.8%) of the 16 male short-tailed cane mice in size classes III and IV, and 12 (66.7%) of the 18 female short-tailed cane mice in size classes III and IV were infected with GTOV ( Table 2). The prevalence of infection in the 34 short-tailed cane mice in size classes III and IV differed signifi cantly from that in the 30 cane mice in size classes I and II (2-tailed Fisher exact test p<0.01), but the prevalence of infection in the 16 male short-tailed cane mice in size classes III and IV did not differ signifi cantly from that in the 18 female short-tailed cane mice in size classes III and IV (2-tailed Fisher exact test p = 0.72).
Thirteen (86.7%) of the 15 female short-tailed cane mice in size class II, 7 (53.8%) of the 13 female short-tailed cane mice in size class III, and 5 (100%) of the 5 female shorttailed cane mice in size class IV were pregnant. Furthermore, 4 (30.8%) of the 13 pregnant short-tailed cane mice in size class II, 5 (71.4%) of the 7 pregnant short-tailed cane mice in size class III, and all 5 of the pregnant short-tailed cane mice in size class IV were infected with GTOV.
As indicated previously, 12 (66.7%) of the 18 female short-tailed cane mice in size classes III and IV were infected with GTOV. Yet only 1 (14.3%) of the 7 shorttailed cane mice in size class I was infected with GTOV ( Table 2). The difference between the prevalence of infection in the female short-tailed cane mice in size classes III and IV and that in the short-tailed cane mice in size class I was signifi cant (2-tailed Fisher exact test p = 0.03).
Seventeen (65.4%) of the 26 Alston's cotton rats in size classes I and II, 19 (61.3%) of the 31 Alston's cotton rats in size classes III and IV, 9 (50.0%) of the 18 male Alston's cotton rats in size classes III and IV, and 10 (76.9%) of the 13 female Alston's cotton rats in size classes III and IV were infected with PIRV ( Table 2). The prevalence of PIRV infection in the Alston's cotton rats in size classes III and IV did not differ signifi cantly from that in the Alston's cotton rats in size classes I and II (2-tailed Fisher exact test p = 0.78), and the prevalence of infection in the male Alston's cotton rats in size classes III and IV did not differ signifi cantly from that in the female Alston's cotton rats in size classes III and IV (2-tailed Fisher exact test p = 0.16).

Discussion
The results of this study affi rm conclusions drawn from previous studies (2,5). Specifi cally, the short-tailed cane mouse is the principal host of GTOV, and Alston's cotton rat is the principal host of PIRV. Examples of GTOV infection in rodents other than the short-tailed cane mouse are limited to the isolation of AV 97021116 from Alston's cotton rat FHV-4149 and the isolation of GTOV from a pygmy rice rat (Oligoryzomys sp.) and 4 Alston's cotton rats captured at localities in Venezuela other than Hato Maporal (11). Similarly, examples of PIRV infection in rodents other than Alston's cotton rats are limited to the isolation of VAV-956 from a short-tailed cane mouse captured on Hato Maporal in 1994 (5) and the isolation of PIRV from 5 short-tailed cane mice and a spiny rat (Proechimys sp.) captured at other localities in Venezuela (12).
The results of the analysis of the capture data in this study suggest that the short-tailed cane mice captured on sites A and B lived in close physical association with Alton's cotton rats and vice versa. Yet none of the 57 shorttailed cane mice captured on A or B were infected with PIRV, and only 1 of the 55 Alston's cotton rats captured on A or B was infected with GTOV. Collectively, these observations suggest that intimate social interactions between short-tailed cane mice and Alston's cotton rats are infrequent. Alternatively, GTOV-infected short-tailed cane mice are rarely infectious to Alston's cotton rats and PIRV-infected Alston's cotton rats are rarely infectious to short-tailed cane mice.
Chronic infections in individual rodents appear to be critical to the long-term maintenance of arenaviruses in nature. Factors that likely affect the duration of GTOV infection in naturally infected short-tailed cane mice include age at exposure to GTOV, host genetics, virus genetics, inoculum dose, and route of exposure (13,14).
The positive association between prevalence of infection and size class in the short-tailed cane mice suggests  that most GTOV infections in short-tailed cane mice are acquired in an age-dependent manner. Allogrooming, mating, intraspecies aggression, and other activities that entail close physical contact may facilitate horizontal transmission in Z. brevicauda mice. The isolation of GTOV from the samples of lung but not the samples of spleen or kidney from 3 antibody-negative short-tailed cane mice (Table 4) suggests that these animals were infected by way of the respiratory tract rather than by wounding or venereal contact. The lack of an association between prevalence of infection and sex in the short-tailed cane mice in size classes III and IV suggests that male animals and female animals contribute equally to the transmission of GTOV in Z. brevicauda mice.
Under the assumption that short-tailed cane mice whelp their fi rst offspring after they reach size class III, the high prevalence of infection in the female cane mice in size classes III and IV together with the low prevalence of infection in the cane mice in size class I suggest that vertical (dam-to-progeny) transmission of GTOV among Z. brevicauda mice is uncommon. Perhaps GTOV infection in the cane mouse fetus is lethal late in gestation. Alternatively, the survivorship of congenitally infected short-tailed cane mice may be signifi cantly less than the survivorship of their uninfected counterparts during birth through weaning.
Together, the high prevalence of PIRV infection in the Alston's cotton rats in size class 1 and the lack of an association between prevalence of infection and size class in the Alston's cotton rats suggest that most cotton rats become infected with PIRV at an early age, perhaps in utero or immediately postpartum. Hypothetically, vertical (dam-to-progeny) virus transmission is the dominant mode of PIRV transmission in S. alstoni rats.
Arenavirus was isolated from the throat swabs and/or samples of urine from 20 (64.5%) of the 31 infected shorttailed cane mice and 30 (83.3%) of the 36 PIRV-infected Alston's cotton rats in this study ( Table 4), suggesting that bodily secretions or excretions from most GTOVinfected short-tailed cane mice and most PIRV-infected Alston's cotton rats can transmit the viruses to humans. In a laboratory study (13), newborn, juvenile, and some adult short-tailed cane mice inoculated with GTOV strain INH-95551 persistently shed virus in saliva, respiratory secretions, or urine through day 208 postinoculation. Whether the magnitude and duration of virus shedding in PIRV-infected Alston's cotton rats are comparable with that in GTOV-infected short-tailed cane mice has not been investigated.
VHF was fi rst recognized as a distinct clinical entity during an outbreak of hemorrhagic fever that began in 1989 in Guanarito (1) GTOV virus is presumed to be the only agent of VHF; however, the majority of the arenaviruses isolated from VHF patients during September 1989 through December 2006 were identifi ed as strains of GTOV solely on the basis of the results of an IFAT in which extensive crossreactivity between PIRV and GTOV is possible (5). Thus, the arenaviruses isolated from some VHF cases may be strains of PIRV.
PIRV, in association with S. alstoni rats, is widely distributed in rural areas in Guanarito and elsewhere in Portuguesa State (2,5,12). Furthermore, Alston's cotton rats (like short-tailed cane mice) are common in grassdominated habitats, for example, tall grass and hedgerows adjacent to cultivated fi elds and areas with tall grass alongside human dwellings. Thus, the epidemiology of PIRV infection likely is highly similar to the epidemiology of GTOV infection in Portuguesa, with most infections in persons who live or work in rural areas in Guanarito.
The neighbor-joining analysis of N protein sequence data separated the viruses from Hato Maporal into 4 groups: GTOV-1, GTOV-2, PIRV-1, and PIRV-2 (Figure). Each group included a strain from 1994 and strains from 1997, suggesting that multiple evolutionary lineages of GTOV and multiple evolutionary lineages of PIRV were maintained on Hato Maporal during mid-June 1994 through early February 1997. Whether GTOV-1 viruses differ from GTOV-2 viruses with regard to pathogenicity in humans or cane mice and whether PIRV-1 viruses differ from PIRV-2 viruses with regard to pathogenicity in humans or cotton rats has not been investigated.