Airbome Transmission of Polyoma Virus 1, 2 Gerard J. McGarrity, Lewis L. Coriell, and Victoria Ammen 3, SUMMARY-Polyoma virus (PV) infection was transmitted through the air of an animal laboratory. Mice free of detectable antibodies to PV were exposed for 1 month to the airborne environment of laboratories housing naturally infected mice. The seroconversion rate was 75% (24/32), as measured by hemagglutination inhibition. Control mice, housed in the sterile atmosphere of a mass air f10w cabinet (MAFC) in the same laboratory, had a seroconversion rate of 15.8% (3/19). Airborne transmission occurred via PV aerosols, generated by the handling of contaminated bedding, cages, and mice during weekly housekeeping. Length of exposure to PV aerosols correlated with seroconversion. One- and 3-hour exposures resulted in seroconversion rates of 40% (6/15) and 72% (23/32), respectively. Seroconversion rates of mice continuously housed in MAFC totaled 5% (2/40). Checkerboarding mice free of detectable antibodies with mice given 10 5 mean tissue culture infective doses of PV ip resulted in an airborne infection rate of 50% (15/30) in a conventionally ventilated room during a 12-week study. The airborne transmission rate was 10% (3/30) when experiments were performed in a mass air f10w room with a vertical air velocity of 26 feet/minute, and 0% (0/30) when a vertical air velocity of30 feet/minute was used. Antibodies to PV could not be detected in any of 138 human sera, including sera from 29 animal-care technicians who handled PY-infected mice and 15 personnel who had worked with the virus.-J Natl Cancer Inst 56: 159-162, 1976.
Polyoma virus (PV) is a frequently encountered virus of rodents (I~ 2) and is present in a high percentage of mouse colonies (3, 4). It is oncogenic for some rodents (5). PV is excreted in the saliva, urine, and feces of infected animals and has been recovered from bedding (5). Infection between cage mates spreads slowly (5). Hemagglutination inhibition (HI) antibody develops and persists for prolonged periods (6). With the use of HI antibodies as an index of infection, studies have reported higher animal infection rates when PV is used in the laboratory. The routes of infection have not been completely c1arified. The influence of fomites (5, 7) and the possibility of airborne spread indicated by cross infection between cages (6, 8) have been discussed. Two other murine viruses, Sendai and Reo-3, can be transmitted through the air (9~ 10). Our objective was to determine if PV is transmitted through the air and to study preventive measures. MATERIALS AND METHODS
Animals.-Specific-pathogen-free mice, strain NLW, were obtained from National Animal Laboratories, O'Fallon, Missouri. Axenic mice were obtained from Charles River Breeding Laboratories, North Wilmington, Massachusetts. Serology.-HI antibodies to PV were assayed with procedures outlined by Parker et al. (11) except that 8 hemagglutination units, not 16, were used. This substitution was effective in controlled studies (12). A titer of 1: 20 or greater was considered positive. Guinea pig erythrocytes from animals screened for maximum agglu-
4
tination by PV were used. Sera from experimental and control animals were coded blindly. Virus antigens, control antigens, and specific hyperimmune antisera were obtained from Microbiological Associates, Inc., Bethesda, Maryland. HI tests inc1uded antigen controls, positive antisera, and negative antisera from axenic mice. Airborne infection studies.-Two types of studies were done, one with mice naturally infected with PV and one with inoculated mice. NLW mice free of detectable antibodies to PV in HI tests in this laboratory were used as sentineis of airborne PV infection. Sentinel mice were exposed to airborne, but not contact, infection in an animal laboratory containing either naturally infected or artificially infected mice in separate experiments. Aerosols of PV from the bedding of infected mice were generated by personnel during housekeeping, cage changing, and animal handling. All sentinel mice were handled wi th sterile forceps by one technician wearing gown, gloves, hair covering, and mask. Gloves were changed between cages. Ini tially cages, bedding, water, and food were sterilized, but studies showed this was unnecessary, Sentinel mice were orbitally bled weekly and their sera assayed for antibodies to PV. As a control on airborne transmission and to study preventive measures, additional sentinel mice handled as above and exposed to infected mice were housed in sterile atmospheres provided by mass air flow. A conventionally ventilated room (CVR) denotes animal laboratories that had a ventilation rate of 15 air changes/hour through medium efficiency filters. Naturally infected mice.-These studies, lasting 4 weeks, were done in a CVR that housed approximately 200 C57BL/6 mice. Twenty-five of these mice were monitored for HI antibodies against PV and 40% were positive. One group of sentinel mice were housed in shoebox cages, 1 mouse/cage. Cages with sentinel mice were placed on the top of solid-shelf animal racks, 6 Ieet above the floor. A second group of sentinel mice were housed in a mass air flow cabinet (MAFC) in the CVR. The MAFC was equipped with a high-efficiency particulate air (HEPA) filter and a uniform vertical air velocity of 35 feetyminute (fpm). Details on this MAFC, which maintained the animals in a sterile atmosphere, are in (10). Artificially injected mice.-These studies, lasting 12 weeks, were done in a CVR and a mass air flow room (MAFR) that were of similar size. NLW mice were inoculated ip with 105 mean tissue culture infective doses (TCID50) of PV in a volume of 0.1 ml. Cages of infected mice (3jcage) were checkerboarded with cages of sentine1 mice (l/cage). Two racks of cages were housed Received May 27, 1975; accepted August 18, 1975. Supported in part by Public Health Service grants RR00903 and RR05582 from the Division of Research Resources, and grantin-aidM-43 from the State of New Jersey. 3 Department of Microbiology, Institute for Medical Research, Camden, N.J. 08103. 4 We thank Riley Hansom, Judith Gager, and CarIene O'Keefe for expert technical assistance. 1
2
JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 56, NO. 1, JANUARY 1976
159
160
MCGARRITY t CORIELL, AND AMME:N
per room with a total of 30 cagesjrack, 15 of each group. There were 30 sentinel micejroom. Animal racks with perforated shelves and shoebox cages with fine-wire screen tops were used. These minimized obstructions to air ßow in the MAFR and prevented the sedimentation of contaminated bedding and other large particles into cages. The MAFR had an entire. ceiling of HEPA filters and a vertical air velocity that could be varied from 10 to 100 fpm. In these studies, 26 and 30 fpm, which effected 223 and 257 air changesjhour, respectively, were used. These velocities were lower than the 100 fpm used in so-called laminar air ßow (13).
10
10
l.-Effect of length of exposure to PV aerosols by naturally infected mice on seroconversion of NLW mice.
TEXT-FlGURE
40 IfMl'
•
COIt"....
Io.... "00.
OMA' "00'" 10
HOUR.
IX~OauRe
ence between 1- and 3-hour exposures in the control groups was also significant at the 0.05 probability level.
RESULTS
Infection With Naturally Infected Mice
Table 1 lists results of studies on the airborne spread of PV by naturally infected animals. Of the sentinel mice in the CVR, 75% developed antibody to PV compared to 15.8% in the MAFC. Two of 3animals in the MAFC that seroconverted were in the NIAFC when the motors were turned off for approximately 1 hour because of mechanical difficulty. The difference between the MAFC and the CVR is statistically significant at the 0.01 probability level. Eight of 10 animals housed in the MAFC with the motors turned off for 4 weeks developed antibodies to PV. Infection Transmitted During Activity
To determine if infection was transmitted by aerosols generated during the transfer of animals from dirty to clean cages, sentinel mice were housed in an MAFR having a velocity of 26 fpm. Some animals were removed and housed in the CVR for varying periods, but only during animal transfer. Transfer required approximately 1 hour jweek. In separate sets of experiments lasting 4 weeks, sentinel mice were subjected to either one or three weekly exposures to the aerosols generated during housekeeping. Results of these studies are in text-figure 1. One-hour exposure resulted in a 40% (6 j 15) seroconversion. The average titer of the 6 seroconverted animals was 1: 60; two developed titers of I: 20; three had titers of 1: 40; and one had a titer of 1: 80. Five of 6 positive animals had detectable titers 1 week after exposure; antibodies in the sixth animal were first detected 2 weeks after exposure. Five of 6 seroconverters had detectable antibodies for at least 2 weeks; the remaining animal (titer 1: 20) was positive for 1 week only. After three weekly l-hour exposures, 72% {23j32) of the mice developed titers. The average titer in the seropositive animals was 1: 80, and antibodies persisted for a minimum of 2 weeks. The difference between experimental and control groups for both exposures in text-figure 1 was statistically significant at the 0.01 probability level. The differ-
Infection With Artificially Infected Mice
In studies with artificiaUy infected animals, all mice inoculated with 105 TCID50 of PV developed antibodies. The average titer of 60 mice was 1: 4,100. In additional control studies, 102 . 5 TCID50 of PV given ip was sufficient to initiate response in all mice. The inoculation of 101.5 and 10°·5 TCID50 of PV ip elicited an antibody response in 95 and 85%, respectively. An average titer of 1: 1,824 developed in mice receiving this lowest inoculum. Mice receiving 105 TCID50 of PV had a me an titer of 1: 320 fourteen weeks. after inoculation, and had 102-10 4 TCID50 of PV virus per ml of urine 1 month after inoculation. The cages of PV-infected and sentinel mice were checkerboarded; the results are shown in text-figure 2. Of the sentinel mice in the CVR, 50% developed antibodies to PV during the 12-week study. In the MAFR, with a velocity of. 26 fpm, 10% were infected. There were no infections in the MAFR with a velocity of 30 fpm. The differences between the control and 26 fpm and 30 fpm are statistically significant at the 0.01 proba-
60
• 50 Z
,x--x Control
.!
r/
~
e
~ 40
,;
c
0
.;; 30 ~
•>
c ~
/
_-x--x
•
20
0
t /)
TABLE
l.-Airborne polyoma infection of sentinel mice by naturally infected mice
CVR 24/3~
(75%)
MAFCa
~
10 0
On
Off
3/19 (15.8%)
8/10 (80%)
"Inside CVR; filter system was turned on or off in separate expts; values are number infected/No. exposed (%) in a 4-wk study.
/
/
I
I
/
/'
,,°'--0---0---0---0
26 ,
pm
I / ;Of----o.--
;_,_;,' .~_._._._._._._._.306pm
2
4
6 WEEKS
2.-Airborne PV infection rates in artificially infected mice housed in a CVR, and an MAFR at 26 and 30 fpm.
TEXT-FIGURE
AIRBORNE POLYOMA VIRUS INFECTION TABLE
2.-Survey of human sera fOT antibodies to PV Personnel a
Animal-care
Administrative
Laboratory
Total
0/29 (41)
0/31 (53)
0/78 (130)
0/138 (224)
4 Number positive/No. tested; numbers in parentheses denote total number of tests.
bility level. The differences in results at 26 fpm and 30 fpm are not significant at the 0.05 probability level. Mice that seroconverted in the CVR developed an average titer of 105 (range, 20-1,280). All were positive a minimum of 2 weeks. PV was recovered from the urine of 2 of 6 seroconverted mice, as measured by the mouse antibody production test. The amount of virus was not quantitated. Limited environmental samplings of the bedding, dust, and floor failed to detect the virus. Air sampling was initially negative, but PV was eventually recovered from the air with a large-volume air sampler, Bacterial air tests with Andersen sam plers showed that aerosols of between 55 and 135 colony-forming units I cubic foot of air were generated during animal transfer. Between 50 and 80% of viable airborne particulates were larger than 9.2 JL. Human Seroconversion
Human sera, obtained over a l-year period from animal-care, cell-culture, and administrative personnel were tested for antibodies to PV (table 2). No HI antibodies were detected in any of the 138 employees tested, including 29 animal-care personnel who handled PV-infected animals. At least 15 cell-culture employees had worked with the virus in the previous 10 years, and 3 had worked extensively with it the past 2 years. Of these 138 personneI, 86 were tested twice, 12 months apart. DISCUSSION
PV is excreted in copious amounts in the urine of infeeted mice as long as 9 months after inoculation (14). It is excreted in lesser amounts in the feces and saliva and has been recovered from the bedding of cages with infected mice (5). Brodsky et al. (7) and Eddy et al. (15) demonstrated the resistance of PV to environmental influences. Akers et al. (16) showed that another papovavirus, simian virus 40, was stable in aerosol over a broad relative humidity range, 15-90% at 21 0 C. The excretion of large amounts of virus into the bedding and its relative resistance to environmental inactivation make it a candidate for airborne transmission. Microbiologic aerosols are generated from contaminated bedding during cage handling (17, 18). The bacterial particles (50-80%) in these aerosols of our study were larger than 9.2 JL. These sediment 1 fpm or faster; the remainder will stay airborne longer. Airborne PV probably does not exist as single particles but rides larger particles of hair, bedding, and excreta. Excretion of microorganisms into the bedding and subsequent airborne spread are among the factors that should be considered in animal laboratory facilities. Airborne transmission can readily occur during the generation of infectious aerosols, i.e., handling of bedding and animals. Such airborne spread has been documented for Sendai, Reo-3 viruses, and Proteus mirabilis (9, 10, 17). Parker and Reynolds (19) conc1uded airborne transmission was of little importance in Sendai virus infection in
161
mice, but in this study animals were housed on bare wire without bedding. Bedding, cage spacing, and the number of animals per cage influence airborne infection (17,18). I t is not known whether the mice in this study were infected by inhalation or by ingestion of infective partieIes. The average tidal volume of mice is 0.15 ml, respiratory frequency is 163 per minute, and the minute volurne is 0.024 liters (20). Mice inhale an average of 1.44 liters of air jhour. By extrapolation of these data and those presented in text-figure I, where 40% of the animals were infected after a l-hour exposure, the miee in our studies were exposed to one mean infectious dose (ID50) of PV in 75 minutes, i.e., after breathing 1.8 liters of air. When 106 .4 ID50 of PV was fed ip daily for 3 successive days to weanlings, fewer than half developed antibody (14); this suggested that ingestion is not an important mode of transmission. In both weanlings and newborns, 102-103 ID50 were required to infect by intranasal inoculation (14). This concentration is within the range of virus excreted in the urine. Inhalation has been suggested as a portal of entry (8). In preliminary studies, we isolated PV from the air wth a large-volume air sam pler, The details of sampling will be published separately. For 12 months no HI antibodies were demonstrated in 138 personnel, inc1uding 29 animal-care technicians and 15 other persons who worked with the virus. This may reflect the effectiveness of microbiologic barriers, the poor humoral immunogenicity of PV for humans, or the transitory nature of antibody response to the relatively low doses, as seen in some mice in these studies. PV does not reproduce in human cells (21). Stewart et al. reported seroconversion in a laboratory worker (22), but this was based on one seroconversion from two laboratory workers tested. Our observations suggest that this is rare. More data are needed on the serologie responses of laboratory personnel to other oncogenic viruses. The ip inoculation of mice with PV yields high antibody titers; however, low titers, generally 160 or less, were detected in seroconversions after environmental exposure. The sensitivity of these HI tests was increased by the use of 8 hemagglutination unitsjtest and by the selection of guinea pig erythrocytes that were maximally sensitive to agglutinatiOl). by PV. A double-blind coding system, changed weekly, was used. Positive sera generally remained positive for at least 2 weeks. The low titers may reflect the small infectious dose in these experiments compared to laboratory inoculation. Rowe (5) stated that the dose of PV acquired spontaneously is probably much smaller than that used in tumorigenie experiments. Yabe et al. (3) noted that uninoculated mice positive for antibodies to PV had lower titers than did mice with induced tumors. The lower incidence of seroconversion observed with exposure to artificially versus naturally infected mice may be due to several differences in animal housing, ineIuding type of cage top, cage spacing, animal rack, and animal density. The MAFC and MAFR used here significantly reduced or eliminated the incidence of PV infection. These systems differ from conventional laminar air-flow systems that use velocities of 100 fpm. Other controlled studies demonstrate the effectiveness of MAFC and MAFR systems with other viruses and bacteria (10, 17, 18). The only other observed difference between the CVR and the MAFR in this study was humidity (55%
162
MCGARRITY, CORIELL, AND AMMEN
relative humidity in the MAFR and 75% in the CVR). This does not affeet the stability of aerosolized papovavirus (16). Other ways to reduee airborne infeetion in an animal eolony include quarantine, isolation, good housekeeping, and supervision of personnel. Filter-top eages have been reeommended (23J 24), but C02 and NH3 ean inerease inside (25). Drug metabolism by hepatic mierosomes deereases in animals housed in filter-top eages and dirty environments (26). An inerease in turbulent ventilation should theoretieally reduee airborne eontamination, but in one study, 38 air ehangesjhour effeeted only small reduetions (18). Optimal methods to implement mass air ßow systems in animal facilities are under study, induding eage spaeing, effeet of obstruetion to air ßow, ßow rates, eost of operation, animal density, and physiologie and environmental effeets.
REFERENCES (1) GROSS L: A filterable agent recovered from AK leukemic extracts, causing salivary gland careinomas in C3H mice, Proc Soc Exp Biol Med 83:414-431, 1953 (2) STEWART SE: Leukemia in mice produced by a filterable agent present in AKR leukemic tissues, with notes on a sarcoma produced by the same agent. Anat Rec 117:532, 1953 (3) YABE Y, NERUSHI S, SATO Y, et al: Distribution of hemagglutination-inhibiting antibodies against polyoma virus in laboratory mice. J Natl Cancer Inst 26:621-628, 1961 (4) PARKER JC, TENNANT RW, WARD TG: Prevalence of viruses in mouse colonies. Natl Cancer Inst Monogr 20:25-36, 1966 (5) ROWE WP: The epidemiology of mouse polyoma infection. Bacteriol Rev 25:18-31, 1961 (6) ROWE·WP, HARTLEY JW, LAW LW, et al: Studies of mouse polyoma virus infection. 111. Distribution of antibodies in laboratory mouse colonies. J Exp Med 109:449-462, 1959 (7) BRODSKY I, ROWE WP, HARTLEY JW, et al: Studies of mouse polyoma virus infection. 11. Virus stability. J Exp Med 109:439-447, 1959 (8) ROWE WP, HARTLEY JW, BRODSKY I, et al: Observations on the spread of mouse polyoma virus infection. Nature 182:1617, 1958 (9) VAN DER VEEN J, POORT Y, BIRCHFIELD DJ: Experimental transmission of Sendai virus infeetion in mice. Areh Gesamte Virusforsch 31:237-246, 1970 (10) MCGARRITY GJ, CORlELL LL: Mass airßow cabinet for control of airborne infeetion of laboratory rodents. Appl Microbiol 26:167-172, 1973
(11) PARKER JC, TENNANT RW, WARD TG, et al: Virus studies with germfree mice, I. Preparation of serologie diagnostic reagents and survey of germfree and monocontaminated mice for indigenous murine viruses. J Natl Cancer Inst 34:371380, 1965 (12) HIERHOLZER JC, SUGGS MT, HALL EC: Standardized viral hemagglutination and hemagglutination inhibition tests. II. Deseription and statistieal evaluation. Appl Microbiol 18:824-833, 1969 (lJ) CORIEl.L LL, MCGARRITY GJ: Biohazard hood to prevent infection during microbiological procedures. Appl Microbiol 16:1895-1900, 1968 (14) ROWE WP, HUEBNER RJ, HARTLEY JW: Ecology of a mouse tumor virus. In Perspectives in Virology (Pollard M, ed.). Minneapolis, Burgess Publishing Co. 1961, pp 177-190 (15) EDDyBE, STEWART SE, GRUBBS GE: Inßuence of tissue culture passage, storage, temperature and drying on viability of SE polyoma virus. Proc Soc Exp Biol Med 99:289-292, 1958 (16) AKERS TG, PRATO CM, DUBOVI EJ: Airborne stability of simian virus 40. Appl Microbiol 26: 146-148, 1973 (17) MCGARRITY GJ, CORIELL LL, SCHAEDLER RW, et al: Medical applications of dust-free rooms. 111. Use in an animal care laboratory. Appl Microbiol 18:142-146, 1969 (18) - - - : Studies on airborne infection in an animal care laboratory. In Development in Industrial Microbiology (Corum C, ed.). Washington, D.C., American Institute for Biological Sciences, 1970, pp 58-64 (19) PARKER JC, REYNOLDS RK: Naturalhistory of Sendai virus infeetion in mice. Am J Epidemiol 88:112-125, 1968 (20) GUYTON AC: Measurements of respiratory volumes of laboratory animals. Am J Physiol 150:70-77, 1947 (21) GRVEN R, GRAESSMANN M, GRAESSMAN A, et al: Infection of human cells with polyoma virus. Virology 58:290-293, 1974 (22) STEWART SE, EDDY BE, STANTON MF: Induction of neoplasma in mice and other mammals by a tumor agent carried in tissue culture. Proc Can Cancer Conf 3:287-305, 1959 (23) KRAFT LM: Observations on the' control and natural history of epidemie diarrhea of infant mice (EDIM). Yale J Biol Med 31: 121-137, 1958 (24) SCHNEIDER HA, COLLINS GR: Successful prevention of infantile diarrhea of mice during an epizootie by means of a new filter cage unopened from birth to weanling. Lab Anim Sei 16:60-71, 1966 (25) SERRANO LJ: Carbon dioxide and ammonia in mouse cages; effect of cage covers, population and activity. Lab Anim Sei 21:75-85, 1971 (26) VESELL ES, LANG CM, WHln: WJ, et al: Hepatic drug metabolism in rats: Impairment in a dirty environment. Science 179:896-897, 1973
Polyoma virus (PV) is a frequently encountered virus of rodents (I~ 2) and is present in a high percentage of mouse colonies (3, 4). It is oncogenic for some rodents (5). PV is excreted in the saliva, urine, and feces of infected animals and has been recovered from bedding (5). Infection between cage mates spreads slowly (5). Hemagglutination inhibition (HI) antibody develops and persists for prolonged periods (6). With the use of HI antibodies as an index of infection, studies have reported higher animal infection rates when PV is used in the laboratory. The routes of infection have not been completely c1arified. The influence of fomites (5, 7) and the possibility of airborne spread indicated by cross infection between cages (6, 8) have been discussed. Two other murine viruses, Sendai and Reo-3, can be transmitted through the air (9~ 10). Our objective was to determine if PV is transmitted through the air and to study preventive measures. MATERIALS AND METHODS
Animals.-Specific-pathogen-free mice, strain NLW, were obtained from National Animal Laboratories, O'Fallon, Missouri. Axenic mice were obtained from Charles River Breeding Laboratories, North Wilmington, Massachusetts. Serology.-HI antibodies to PV were assayed with procedures outlined by Parker et al. (11) except that 8 hemagglutination units, not 16, were used. This substitution was effective in controlled studies (12). A titer of 1: 20 or greater was considered positive. Guinea pig erythrocytes from animals screened for maximum agglu-
4
tination by PV were used. Sera from experimental and control animals were coded blindly. Virus antigens, control antigens, and specific hyperimmune antisera were obtained from Microbiological Associates, Inc., Bethesda, Maryland. HI tests inc1uded antigen controls, positive antisera, and negative antisera from axenic mice. Airborne infection studies.-Two types of studies were done, one with mice naturally infected with PV and one with inoculated mice. NLW mice free of detectable antibodies to PV in HI tests in this laboratory were used as sentineis of airborne PV infection. Sentinel mice were exposed to airborne, but not contact, infection in an animal laboratory containing either naturally infected or artificially infected mice in separate experiments. Aerosols of PV from the bedding of infected mice were generated by personnel during housekeeping, cage changing, and animal handling. All sentinel mice were handled wi th sterile forceps by one technician wearing gown, gloves, hair covering, and mask. Gloves were changed between cages. Ini tially cages, bedding, water, and food were sterilized, but studies showed this was unnecessary, Sentinel mice were orbitally bled weekly and their sera assayed for antibodies to PV. As a control on airborne transmission and to study preventive measures, additional sentinel mice handled as above and exposed to infected mice were housed in sterile atmospheres provided by mass air flow. A conventionally ventilated room (CVR) denotes animal laboratories that had a ventilation rate of 15 air changes/hour through medium efficiency filters. Naturally infected mice.-These studies, lasting 4 weeks, were done in a CVR that housed approximately 200 C57BL/6 mice. Twenty-five of these mice were monitored for HI antibodies against PV and 40% were positive. One group of sentinel mice were housed in shoebox cages, 1 mouse/cage. Cages with sentinel mice were placed on the top of solid-shelf animal racks, 6 Ieet above the floor. A second group of sentinel mice were housed in a mass air flow cabinet (MAFC) in the CVR. The MAFC was equipped with a high-efficiency particulate air (HEPA) filter and a uniform vertical air velocity of 35 feetyminute (fpm). Details on this MAFC, which maintained the animals in a sterile atmosphere, are in (10). Artificially injected mice.-These studies, lasting 12 weeks, were done in a CVR and a mass air flow room (MAFR) that were of similar size. NLW mice were inoculated ip with 105 mean tissue culture infective doses (TCID50) of PV in a volume of 0.1 ml. Cages of infected mice (3jcage) were checkerboarded with cages of sentine1 mice (l/cage). Two racks of cages were housed Received May 27, 1975; accepted August 18, 1975. Supported in part by Public Health Service grants RR00903 and RR05582 from the Division of Research Resources, and grantin-aidM-43 from the State of New Jersey. 3 Department of Microbiology, Institute for Medical Research, Camden, N.J. 08103. 4 We thank Riley Hansom, Judith Gager, and CarIene O'Keefe for expert technical assistance. 1
2
JOURNAL OF THE NATIONAL CANCER INSTITUTE, VOL. 56, NO. 1, JANUARY 1976
159
160
MCGARRITY t CORIELL, AND AMME:N
per room with a total of 30 cagesjrack, 15 of each group. There were 30 sentinel micejroom. Animal racks with perforated shelves and shoebox cages with fine-wire screen tops were used. These minimized obstructions to air ßow in the MAFR and prevented the sedimentation of contaminated bedding and other large particles into cages. The MAFR had an entire. ceiling of HEPA filters and a vertical air velocity that could be varied from 10 to 100 fpm. In these studies, 26 and 30 fpm, which effected 223 and 257 air changesjhour, respectively, were used. These velocities were lower than the 100 fpm used in so-called laminar air ßow (13).
10
10
l.-Effect of length of exposure to PV aerosols by naturally infected mice on seroconversion of NLW mice.
TEXT-FlGURE
40 IfMl'
•
COIt"....
Io.... "00.
OMA' "00'" 10
HOUR.
IX~OauRe
ence between 1- and 3-hour exposures in the control groups was also significant at the 0.05 probability level.
RESULTS
Infection With Naturally Infected Mice
Table 1 lists results of studies on the airborne spread of PV by naturally infected animals. Of the sentinel mice in the CVR, 75% developed antibody to PV compared to 15.8% in the MAFC. Two of 3animals in the MAFC that seroconverted were in the NIAFC when the motors were turned off for approximately 1 hour because of mechanical difficulty. The difference between the MAFC and the CVR is statistically significant at the 0.01 probability level. Eight of 10 animals housed in the MAFC with the motors turned off for 4 weeks developed antibodies to PV. Infection Transmitted During Activity
To determine if infection was transmitted by aerosols generated during the transfer of animals from dirty to clean cages, sentinel mice were housed in an MAFR having a velocity of 26 fpm. Some animals were removed and housed in the CVR for varying periods, but only during animal transfer. Transfer required approximately 1 hour jweek. In separate sets of experiments lasting 4 weeks, sentinel mice were subjected to either one or three weekly exposures to the aerosols generated during housekeeping. Results of these studies are in text-figure 1. One-hour exposure resulted in a 40% (6 j 15) seroconversion. The average titer of the 6 seroconverted animals was 1: 60; two developed titers of I: 20; three had titers of 1: 40; and one had a titer of 1: 80. Five of 6 positive animals had detectable titers 1 week after exposure; antibodies in the sixth animal were first detected 2 weeks after exposure. Five of 6 seroconverters had detectable antibodies for at least 2 weeks; the remaining animal (titer 1: 20) was positive for 1 week only. After three weekly l-hour exposures, 72% {23j32) of the mice developed titers. The average titer in the seropositive animals was 1: 80, and antibodies persisted for a minimum of 2 weeks. The difference between experimental and control groups for both exposures in text-figure 1 was statistically significant at the 0.01 probability level. The differ-
Infection With Artificially Infected Mice
In studies with artificiaUy infected animals, all mice inoculated with 105 TCID50 of PV developed antibodies. The average titer of 60 mice was 1: 4,100. In additional control studies, 102 . 5 TCID50 of PV given ip was sufficient to initiate response in all mice. The inoculation of 101.5 and 10°·5 TCID50 of PV ip elicited an antibody response in 95 and 85%, respectively. An average titer of 1: 1,824 developed in mice receiving this lowest inoculum. Mice receiving 105 TCID50 of PV had a me an titer of 1: 320 fourteen weeks. after inoculation, and had 102-10 4 TCID50 of PV virus per ml of urine 1 month after inoculation. The cages of PV-infected and sentinel mice were checkerboarded; the results are shown in text-figure 2. Of the sentinel mice in the CVR, 50% developed antibodies to PV during the 12-week study. In the MAFR, with a velocity of. 26 fpm, 10% were infected. There were no infections in the MAFR with a velocity of 30 fpm. The differences between the control and 26 fpm and 30 fpm are statistically significant at the 0.01 proba-
60
• 50 Z
,x--x Control
.!
r/
~
e
~ 40
,;
c
0
.;; 30 ~
•>
c ~
/
_-x--x
•
20
0
t /)
TABLE
l.-Airborne polyoma infection of sentinel mice by naturally infected mice
CVR 24/3~
(75%)
MAFCa
~
10 0
On
Off
3/19 (15.8%)
8/10 (80%)
"Inside CVR; filter system was turned on or off in separate expts; values are number infected/No. exposed (%) in a 4-wk study.
/
/
I
I
/
/'
,,°'--0---0---0---0
26 ,
pm
I / ;Of----o.--
;_,_;,' .~_._._._._._._._.306pm
2
4
6 WEEKS
2.-Airborne PV infection rates in artificially infected mice housed in a CVR, and an MAFR at 26 and 30 fpm.
TEXT-FIGURE
AIRBORNE POLYOMA VIRUS INFECTION TABLE
2.-Survey of human sera fOT antibodies to PV Personnel a
Animal-care
Administrative
Laboratory
Total
0/29 (41)
0/31 (53)
0/78 (130)
0/138 (224)
4 Number positive/No. tested; numbers in parentheses denote total number of tests.
bility level. The differences in results at 26 fpm and 30 fpm are not significant at the 0.05 probability level. Mice that seroconverted in the CVR developed an average titer of 105 (range, 20-1,280). All were positive a minimum of 2 weeks. PV was recovered from the urine of 2 of 6 seroconverted mice, as measured by the mouse antibody production test. The amount of virus was not quantitated. Limited environmental samplings of the bedding, dust, and floor failed to detect the virus. Air sampling was initially negative, but PV was eventually recovered from the air with a large-volume air sampler, Bacterial air tests with Andersen sam plers showed that aerosols of between 55 and 135 colony-forming units I cubic foot of air were generated during animal transfer. Between 50 and 80% of viable airborne particulates were larger than 9.2 JL. Human Seroconversion
Human sera, obtained over a l-year period from animal-care, cell-culture, and administrative personnel were tested for antibodies to PV (table 2). No HI antibodies were detected in any of the 138 employees tested, including 29 animal-care personnel who handled PV-infected animals. At least 15 cell-culture employees had worked with the virus in the previous 10 years, and 3 had worked extensively with it the past 2 years. Of these 138 personneI, 86 were tested twice, 12 months apart. DISCUSSION
PV is excreted in copious amounts in the urine of infeeted mice as long as 9 months after inoculation (14). It is excreted in lesser amounts in the feces and saliva and has been recovered from the bedding of cages with infected mice (5). Brodsky et al. (7) and Eddy et al. (15) demonstrated the resistance of PV to environmental influences. Akers et al. (16) showed that another papovavirus, simian virus 40, was stable in aerosol over a broad relative humidity range, 15-90% at 21 0 C. The excretion of large amounts of virus into the bedding and its relative resistance to environmental inactivation make it a candidate for airborne transmission. Microbiologic aerosols are generated from contaminated bedding during cage handling (17, 18). The bacterial particles (50-80%) in these aerosols of our study were larger than 9.2 JL. These sediment 1 fpm or faster; the remainder will stay airborne longer. Airborne PV probably does not exist as single particles but rides larger particles of hair, bedding, and excreta. Excretion of microorganisms into the bedding and subsequent airborne spread are among the factors that should be considered in animal laboratory facilities. Airborne transmission can readily occur during the generation of infectious aerosols, i.e., handling of bedding and animals. Such airborne spread has been documented for Sendai, Reo-3 viruses, and Proteus mirabilis (9, 10, 17). Parker and Reynolds (19) conc1uded airborne transmission was of little importance in Sendai virus infection in
161
mice, but in this study animals were housed on bare wire without bedding. Bedding, cage spacing, and the number of animals per cage influence airborne infection (17,18). I t is not known whether the mice in this study were infected by inhalation or by ingestion of infective partieIes. The average tidal volume of mice is 0.15 ml, respiratory frequency is 163 per minute, and the minute volurne is 0.024 liters (20). Mice inhale an average of 1.44 liters of air jhour. By extrapolation of these data and those presented in text-figure I, where 40% of the animals were infected after a l-hour exposure, the miee in our studies were exposed to one mean infectious dose (ID50) of PV in 75 minutes, i.e., after breathing 1.8 liters of air. When 106 .4 ID50 of PV was fed ip daily for 3 successive days to weanlings, fewer than half developed antibody (14); this suggested that ingestion is not an important mode of transmission. In both weanlings and newborns, 102-103 ID50 were required to infect by intranasal inoculation (14). This concentration is within the range of virus excreted in the urine. Inhalation has been suggested as a portal of entry (8). In preliminary studies, we isolated PV from the air wth a large-volume air sam pler, The details of sampling will be published separately. For 12 months no HI antibodies were demonstrated in 138 personnel, inc1uding 29 animal-care technicians and 15 other persons who worked with the virus. This may reflect the effectiveness of microbiologic barriers, the poor humoral immunogenicity of PV for humans, or the transitory nature of antibody response to the relatively low doses, as seen in some mice in these studies. PV does not reproduce in human cells (21). Stewart et al. reported seroconversion in a laboratory worker (22), but this was based on one seroconversion from two laboratory workers tested. Our observations suggest that this is rare. More data are needed on the serologie responses of laboratory personnel to other oncogenic viruses. The ip inoculation of mice with PV yields high antibody titers; however, low titers, generally 160 or less, were detected in seroconversions after environmental exposure. The sensitivity of these HI tests was increased by the use of 8 hemagglutination unitsjtest and by the selection of guinea pig erythrocytes that were maximally sensitive to agglutinatiOl). by PV. A double-blind coding system, changed weekly, was used. Positive sera generally remained positive for at least 2 weeks. The low titers may reflect the small infectious dose in these experiments compared to laboratory inoculation. Rowe (5) stated that the dose of PV acquired spontaneously is probably much smaller than that used in tumorigenie experiments. Yabe et al. (3) noted that uninoculated mice positive for antibodies to PV had lower titers than did mice with induced tumors. The lower incidence of seroconversion observed with exposure to artificially versus naturally infected mice may be due to several differences in animal housing, ineIuding type of cage top, cage spacing, animal rack, and animal density. The MAFC and MAFR used here significantly reduced or eliminated the incidence of PV infection. These systems differ from conventional laminar air-flow systems that use velocities of 100 fpm. Other controlled studies demonstrate the effectiveness of MAFC and MAFR systems with other viruses and bacteria (10, 17, 18). The only other observed difference between the CVR and the MAFR in this study was humidity (55%
162
MCGARRITY, CORIELL, AND AMMEN
relative humidity in the MAFR and 75% in the CVR). This does not affeet the stability of aerosolized papovavirus (16). Other ways to reduee airborne infeetion in an animal eolony include quarantine, isolation, good housekeeping, and supervision of personnel. Filter-top eages have been reeommended (23J 24), but C02 and NH3 ean inerease inside (25). Drug metabolism by hepatic mierosomes deereases in animals housed in filter-top eages and dirty environments (26). An inerease in turbulent ventilation should theoretieally reduee airborne eontamination, but in one study, 38 air ehangesjhour effeeted only small reduetions (18). Optimal methods to implement mass air ßow systems in animal facilities are under study, induding eage spaeing, effeet of obstruetion to air ßow, ßow rates, eost of operation, animal density, and physiologie and environmental effeets.
REFERENCES (1) GROSS L: A filterable agent recovered from AK leukemic extracts, causing salivary gland careinomas in C3H mice, Proc Soc Exp Biol Med 83:414-431, 1953 (2) STEWART SE: Leukemia in mice produced by a filterable agent present in AKR leukemic tissues, with notes on a sarcoma produced by the same agent. Anat Rec 117:532, 1953 (3) YABE Y, NERUSHI S, SATO Y, et al: Distribution of hemagglutination-inhibiting antibodies against polyoma virus in laboratory mice. J Natl Cancer Inst 26:621-628, 1961 (4) PARKER JC, TENNANT RW, WARD TG: Prevalence of viruses in mouse colonies. Natl Cancer Inst Monogr 20:25-36, 1966 (5) ROWE WP: The epidemiology of mouse polyoma infection. Bacteriol Rev 25:18-31, 1961 (6) ROWE·WP, HARTLEY JW, LAW LW, et al: Studies of mouse polyoma virus infection. 111. Distribution of antibodies in laboratory mouse colonies. J Exp Med 109:449-462, 1959 (7) BRODSKY I, ROWE WP, HARTLEY JW, et al: Studies of mouse polyoma virus infection. 11. Virus stability. J Exp Med 109:439-447, 1959 (8) ROWE WP, HARTLEY JW, BRODSKY I, et al: Observations on the spread of mouse polyoma virus infection. Nature 182:1617, 1958 (9) VAN DER VEEN J, POORT Y, BIRCHFIELD DJ: Experimental transmission of Sendai virus infeetion in mice. Areh Gesamte Virusforsch 31:237-246, 1970 (10) MCGARRITY GJ, CORlELL LL: Mass airßow cabinet for control of airborne infeetion of laboratory rodents. Appl Microbiol 26:167-172, 1973
(11) PARKER JC, TENNANT RW, WARD TG, et al: Virus studies with germfree mice, I. Preparation of serologie diagnostic reagents and survey of germfree and monocontaminated mice for indigenous murine viruses. J Natl Cancer Inst 34:371380, 1965 (12) HIERHOLZER JC, SUGGS MT, HALL EC: Standardized viral hemagglutination and hemagglutination inhibition tests. II. Deseription and statistieal evaluation. Appl Microbiol 18:824-833, 1969 (lJ) CORIEl.L LL, MCGARRITY GJ: Biohazard hood to prevent infection during microbiological procedures. Appl Microbiol 16:1895-1900, 1968 (14) ROWE WP, HUEBNER RJ, HARTLEY JW: Ecology of a mouse tumor virus. In Perspectives in Virology (Pollard M, ed.). Minneapolis, Burgess Publishing Co. 1961, pp 177-190 (15) EDDyBE, STEWART SE, GRUBBS GE: Inßuence of tissue culture passage, storage, temperature and drying on viability of SE polyoma virus. Proc Soc Exp Biol Med 99:289-292, 1958 (16) AKERS TG, PRATO CM, DUBOVI EJ: Airborne stability of simian virus 40. Appl Microbiol 26: 146-148, 1973 (17) MCGARRITY GJ, CORIELL LL, SCHAEDLER RW, et al: Medical applications of dust-free rooms. 111. Use in an animal care laboratory. Appl Microbiol 18:142-146, 1969 (18) - - - : Studies on airborne infection in an animal care laboratory. In Development in Industrial Microbiology (Corum C, ed.). Washington, D.C., American Institute for Biological Sciences, 1970, pp 58-64 (19) PARKER JC, REYNOLDS RK: Naturalhistory of Sendai virus infeetion in mice. Am J Epidemiol 88:112-125, 1968 (20) GUYTON AC: Measurements of respiratory volumes of laboratory animals. Am J Physiol 150:70-77, 1947 (21) GRVEN R, GRAESSMANN M, GRAESSMAN A, et al: Infection of human cells with polyoma virus. Virology 58:290-293, 1974 (22) STEWART SE, EDDY BE, STANTON MF: Induction of neoplasma in mice and other mammals by a tumor agent carried in tissue culture. Proc Can Cancer Conf 3:287-305, 1959 (23) KRAFT LM: Observations on the' control and natural history of epidemie diarrhea of infant mice (EDIM). Yale J Biol Med 31: 121-137, 1958 (24) SCHNEIDER HA, COLLINS GR: Successful prevention of infantile diarrhea of mice during an epizootie by means of a new filter cage unopened from birth to weanling. Lab Anim Sei 16:60-71, 1966 (25) SERRANO LJ: Carbon dioxide and ammonia in mouse cages; effect of cage covers, population and activity. Lab Anim Sei 21:75-85, 1971 (26) VESELL ES, LANG CM, WHln: WJ, et al: Hepatic drug metabolism in rats: Impairment in a dirty environment. Science 179:896-897, 1973
