Online Issues
<< All Back-issues
<< This Issue's Table of Contents
ILAR Journal V39(4) 1998
Opportunistic Infections in Rats and Mice
Steven H. Weisbroth, Robert Peters, Lela K. Riley, and William Shek
| Steven H. Weisbroth, D.V.M., is President of AnMed Biosafe Inc., Rockville, Maryland; Robert Peters, Ph.D., is Scientific Director at MA Bioservices, Rockville, Maryland; Lela Riley, Ph.D., is Associate Professor at the University of Missouri, Columbia, Missouri; and William Shek, D.V.M., Ph.D., is Director of Diagnostic Services at Charles River Laboratories, Wilmington, Massachusetts. |
Participants in the September 1997 Workshop on Opportunistic Infections of Rats/Mice, held in Washington, DC, attempted to deal with the issue of defining "opportunistic" agents. Although they were unable to reach consensus regarding universally applicable guidelines to usefully distinguish between "primary pathogens," "opportunists," and "commensals," they did agree that the issue hinges largely on host constitutional factors of relative resistance and susceptibility to given microbial forms.
Impaired immunocompetency was first recognized in the 1950s and 1960s as a physiological consequence of animals undergoing protocols in radiation biology and pharmaceutical corticosteroid and antimetabolite development. Latent, usually clinically silent microbial forms that include Pneumocystis carinii, Corynebacterium kutscheri, Pseudomonas aeruginosa, and Clostridium piliforme became recognized as concomitants of such research (Weisbroth 1995) and were important in prompting development of health surveillance programs. Since then, in addition to chemical modes of impairment of normal rodents, numerous inherited models of immunodeficiency have been developed and used commonly. Examples include athymic (nude) mice and rats and mice with severe combined immunodeficiency (SCID1). Confounding the issue is the proliferating use of mouse and rat inbred strains and transgenics with bioengineered genomes that may have impaired or fragile immunocompetency as a deliberate (or accidental) inherited character. Since the early 1980s, a substantial literature base has accumulated to document disease episodes in immunodeficient rodents with infectious agents that are usually clinically silent in their immunocompetent counterparts.
Workshop participants, for example, considered both "primary pathogen" and "opportunistic" as a definition for defining the pathogen status of Pneumocystis carinii. For immunocompetent rodents, Pneumocystis is a rarely encountered opportunist diagnostically; however, for immunodeficients, there is an established record for Pneumocystis as a primary pathogen (Kitada and Serikawa 1994; Weisbroth 1995). Additionally, a considerable list of viral, protozoan, bacterial, and helminth pathogens are similarly clinically silent in immunocompetent hosts although they produce disease in immunodeficient rodents. A similar situation has been shown with Sendai virus infection in inbred and outbred mouse strains (Parker and others 1978), and considerable data have accumulated more recently on the role of certain Helicobacter species as pathogens of some mouse strains but not others (Ward and others 1994b, 1996). Thus, the Workshop participants' problem was whether to define "pathogen" at the level of the normatively resistant host or at the level of the most susceptible. For the present, their position is to recognize that issues of infectivity and pathogenicity are influenced as much by host-related factors of resistance and susceptibility as by biological properties of specific infectious agents.
Workshop participants acknowledge that health report recipients (investigators and animal facility managers) must be able to understand the significance of positive findings, particularly if clinically inapparent, in making animal colony health management and procurement decisions. They also recognize---and emphasize---that few facile, universally applicable guidelines are available to simplify the process. For example, although there is universal consensus that a sustained finding of antibodies to lymphocytic choriomeningitis virus or isolation of Salmonella typhimurium should trigger prompt measures for containment and eradication, there is no such consensus for findings such as Entamoeba muris, Helicobacter muridarum or Klebsiella oxytoca. It is hoped that the following assessment will help to clarify the issue of opportunistic infections in laboratory rodents.
REVIEW OF DIAGNOSTIC METHODOLOGY
Introduction
Microbiological assessment of the health status of laboratory rats and mice may be defined as the science of evaluating representative sample groups from given units of those species against a specific listing of etiological agents of disease to define the health status of the source colony. The purpose of this information is to prevent introduction of disease and to monitor the microbial status of resident colonies. Development of such data on a repetitive schedule forms the objective basis on which to (1) establish and/or confirm the ongoing microbial status of commercial and institutional rodent production colonies; (2) develop institutional procurement standards for supplier eligibility based on animal health criteria; and (3) continuously monitor the health status of institutional research animal residents, including recent arrivals undergoing equilibration or quarantine before use, those currently involved in research protocols, and those coming off study.
The goal of health surveillance programs is to detect by examination of 1 or more representative sample groups the presence (even in a single individual) of any pathogen from a specific profile of infectious agents. If an agent is detected in the sample group, the larger population (such as the breeding unit or research animal room) should be considered contaminated (or infected) with the same detected agent(s). Of equal importance is the inability to detect any of a specific profile of agents under controlled circumstances. By this process, designated breeding and research units may be demonstrated to be free from a specified list of pathogens, assuming that adequate sample representation and test methodology have taken place.
It is difficult to overstate the additive, compounding effect of repetitive sampling from closed colonies (see Target Testing, below, for additional discussion of sampling statistics). The reliability of the assessment results increases enormously as the total number of animals (or samples) tested from the unit increases with time. Scheduled, repetitive testing also provides current information on a timely basis for continuously updating the health status of recent arrivals at institutions, of sentinelized research colonies, and of breeder production colonies either to detect any changes in the status of such units or to objectively conclude that there have been no changes.
The importance of health surveillance programs is well established (Fox 1977; Jacoby and Barthold 1980; Jacoby and Lindsey 1997; Kraft and others 1994; Loew and Fox 1983; NRC 1991; Rehbinder and others 1996; Small 1984; Waggie and others 1994; Wagner and others 1991). Moreover, the Guide for the Care and Use of Laboratory Animals (NRC 1996), used as the basis for accreditation by the now-international Association for Assessment and Accreditation of Laboratory Animal Care (AAALAC International), recommends a health monitoring component as part of every well-managed animal care program. There is some variation, however, in the way such programs are actually designed, the range and diversity of etiological agents considered of sufficient importance to have on the list, the number of animals (sera or other samples) in the sample group, the periodicity of sample group submission, and the age and sex of animals in the group. Such variation is probably inherent since no single surveillance program could simultaneously meet the needs of all types of institutions. Professional guidance is necessary to shape the surveillance to meet specific institutional needs and to maximize the informational yield from the financial resources available for the purpose. Additional discussion of shaping test panels on the basis of pathogen prevalence appears below.
In shaping the agent profiles against which to evaluate the health status of incoming rodents to prevent introduction of disease and to monitor the status of resident rodent units, animal health specialists and animal users at each institution will need to consider the interaction of specific microorganisms in light of the resident populations' particular rodent strains. These interactions should be considered in evaluating the significance of clinically inapparent pathogens reported as positive findings on sample groups. The significance of such findings is based on precedence in the literature and historical knowledge of likely consequences to hosts of varying immunocompetency. Of course, the likely consequences not only imply direct effects on animal health but also apply equally to the physiological impact of such microbial contamination as a variable in the research process. These considerations lead to interpretation of the acceptability (or nonacceptability) of particular pathogens.
We recommend that animal users resist the tendency to interpret such findings exclusively in limited terms of self-interest. The propensity of microorganisms to spread from cage to cage and room to room by the infectious process is an important and well-documented factor in the considerations above. Experience indicates that the impact of a positive finding (microbial contamination), or information about an animal source with a particular microbe, should be evaluated not only for specific subsets at an institution (that is, a particular cage, room, or investigator's animals) but also more properly in terms of the institutional animal research program as a whole.
Clinical Evaluation, Gross and Microscopic Pathology
Using sampling criteria discussed below, sample groups are selected for diagnostic processing and delivered to the institutional or off-site laboratory. Record keeping should begin with notation of the method and duration of shipment from the source unit because of the obvious effect shipment may have on the clinical appearance and body weight of the submitted animals (Weisbroth and others 1977).
Before processing the shipment group for sample collection, it is important to develop a case history that includes the following identifying particulars of the originating source unit: the premises, on-site location (room number) within the facilities, and strains housed; the genetic and immune status of the sample submission group; individual identification (if any) including sex, approximate age, and pelage color (if haired); and number in the group. The clinical presentation (such as posture, neatness of the hair coat, attitude, presence or absence of discharges from the nares, conjunctiva, and anus) should be noted. However, most often, animals submitted for health surveillance will be clinically normal in appearance if shipped properly. The factors listed above are often helpful in interpretation of results and in establishing the surveillance battery to use for a particular group. For example, it is unnecessary to obtain a blood sample for virus serology from athymic or agammaglobulinemic rodents.
The battery of tests (panel) to be used is determined by careful consideration of the profile of agents meant for detection. Depending on the profile desired, the panels can have reduced or expanded numbers of agents to meet particular screening purposes. In general, agents listed for detection in the profile determine the laboratory tests to be done, and those tests determine which samples should be collected for microbiological evaluation. The sequence of sample collections from various sites of the animal should be carefully structured to prevent (or reduce) sample contamination and to efficiently perform the task.
In Figure 1 appears an example of sample collections for a comprehensive rodent health surveillance screen prototype. As mentioned above, the particulars may differ between laboratories, but most laboratories will follow a sequence similar to that shown in Figure 1.
Diagnostic Detection and Identification of Microorganisms
Identification of potential pathogens in laboratory animals is accomplished using a variety of methodologies. For many years, the "gold standard" for diagnostic identification of microorganisms was culture because growth of the organism from a tissue or biological specimen represented direct evidence of the presence of the organism. Since that time, a number of additional diagnostic techniques have been developed and are now commonly used to identify potential pathogens in laboratory mice and rats. The technique used to identify a specific microorganism depends on a number of factors, including the type of organism, the fastidiousness of the organism, the immunological status of the host, the tissue or site of localization of the organism in the animal, and the particular tests that have been developed to detect and identify the organism.
Culture of bacterial agents is generally accomplished by inoculating medium (such as blood or chocolate agar) with the specimen and incubating the medium in appropriate conditions to allow growth. Morphologically distinct colonies are then evaluated by Gram stain and subjected to a panel of biochemical tests to determine the specific genus and species of each colony type. In certain cases, specimens are cultured on a specialized medium to enhance growth of a potential pathogen or prevent growth of other organisms present in the specimen. For instance, if Helicobacter is suspected, fecal or intestinal specimens are cultured onto blood agar containing antibiotics that limit the growth of intestinal flora, which might otherwise overgrow these slow-growing bacterial species. Although in vitro culture of many bacterial species is easily accomplished, some bacterial species are difficult to grow in vitro because they require special media, environmental conditions, or the presence of mammalian cells. For example, rodent Helicobacters require a microaerophilic atmosphere, and C. piliforme will not grow on cell-free medium but instead requires the presence of certain mammalian cells. For these agents, diagnosis is more difficult, and alternative diagnostic approaches such as polymerase chain reaction (PCR1) or serological assays are often used.
Like bacteria, viral infections can also be diagnosed by in vitro culture. Diagnosis of viral agents can be accomplished by inoculating the clinical specimen into tissue culture. Two types of cultures are used: primary cultures and established immortalized cell lines. Primary cultures are prepared by removing an uninfected organ from the animal, mincing the tissue, and treating with trypsin to disperse the cells into a monolayer. Immortalized cell lines are derived from tumor cells or transformed mammalian cells. In contrast to primary cell lines, which can be passaged only a limited number of times in cell culture, immortalized cell lines can be maintained for many passages. However, most laboratories return to frozen stocks periodically to ensure uniformity. In a diagnostic laboratory setting, 1 or more tissue culture cell lines are usually inoculated with each clinical specimen. Some cell lines are selected on the basis of the tissue tropism of the virus; others are selected because they are known to support growth of a variety of viruses. Cultures are then incubated for several days to weeks and evaluated for the presence of virus by the following methods: detection of characteristic cytopathic effects in the culture such as rounding and detachment of the mammalian cells, formation of multinucleated giant cells, or generation of inclusion bodies. Some viruses require multiple passages before cytopathic effects are evident. Others never cause a visible cytopathic effect but can be recognized by hemagglutination, reactivity of infected cells with virus-specific antibody, or PCR.
Culture of some viruses directly from clinical specimens in tissue culture is particularly difficult. For these viruses, embryonated chicken eggs or explant cultures may be used as alternatives to tissue culture. At one time, embryonated chicken eggs were commonly used for viral propagation. However, with the development of immortalized cell lines, embryonated eggs are now used infrequently for diagnostic evaluations. Explant or organ cultures are similar to primary cell cultures except that the tissue is not treated with trypsin to disperse the cells. Explant cultures allow the virus to proliferate in vitro but obviate the virus' need to adapt to a mammalian cell line, a process that may take many passages for some viruses. Explant cultures have been shown to be particularly useful for initial isolation of some viruses, including rodent parvoviruses (Jacoby and others 1991; Smith and Paturzo 1988). Some viruses cannot be propagated in either mammalian tissue culture systems or embryonated chicken eggs. An example is mouse thymic virus, which must be cultured in vivo in neonatal mice (Osborn 1982). In general, because virus isolation is very labor intensive and often results in false-negative results, it is seldom used as a primary diagnostic test.
Although culture of microorganisms is the standard with which other diagnostic techniques are compared, several other techniques are routinely used to diagnose infectious agents in laboratory animals and biological specimens derived from laboratory animals. Mouse antibody production (MAP1) and rat antibody production (RAP1) tests have been used extensively to identify the presence of infectious virus in murine-derived biological specimens such as ascites or tumors (de Souza and Smith 1989; Morse, 1990; Nicklas and others 1993). In MAP and RAP tests, viral antibody-free mice or rats are inoculated with the specimen, held for several weeks to allow generation of an antibody response to viral agents present in the specimen, and bled. The collected serum is then evaluated serologically for the presence of antibodies to known murine viruses. Although MAP and RAP testing represents an indirect method for virus identification, they are often more sensitive than virus isolation (de Souza and Smith 1989; Morse 1990). The major drawback to antibody production as a diagnostic test is that it is slow (requiring several weeks to yield results) and expensive (since virus-free rodents must be purchased and held in a virus-free environment for several weeks).
Stress or provocation testing is also used to detect murine pathogens. In this approach, the animal is treated with an immunosuppressive drug, such as cyclophosphamide or dexamethasone. This type of testing is based on the premise that rodents infected with a pathogenic microorganism will develop overt disease when suppression of the host immune response allows unchecked growth of the organism. This approach has been used successfully to detect murine pathogens including C. piliforme (Fries 1979; Nakayama and others 1984) and P. carinii (Armstrong and others 1991).
Since the late 1980s, PCR assays have been used to identify murine pathogens. PCR represents a direct diagnostic test since the assay is based on detection of nucleic acid sequences from a particular organism present in a clinical specimen. Briefly, PCR comprises repetitive cycles of a 3-step amplification procedure. In the first step, nucleic acid isolated from a clinical specimen is denatured. In the second step, 2 short oligonucleotide primers complementary to a specific microorganism's genome are allowed to anneal. The oligonucleotide primers are designed based on known sequence and are typically positioned a few hundred nucleotides apart on the genome. In the third step, the oligonucleotide is extended by polymerase. These 3 steps are repeated 20 to 50 times, yielding a product of defined size in reactions containing the microorganism of interest. In contrast, product of the expected size should not be produced from specimens in which the microorganism is not present. To detect microorganisms with RNA genomes, the RNA is first converted to the complementary DNA (cDNA) by the enzyme reverse transcriptase. The cDNA is then used as the template in PCRs as described above.
Major advantages of PCR are its high sensitivity, which typically exceeds that of viral isolation, and its high specificity, which allows differentiation of closely related organisms. The ability of PCR to differentiate closely related species has been exploited in diagnosis of infections in laboratory animals. For example, diagnostic PCR assays have been developed that specifically detect rodent parvovirus species (Besselsen and others 1995a,b) and Helicobacters (Battles and others 1995; Beckwith and others 1997; Riley and others 1996; Shames and others 1995). Another advantage of PCR is that it is rapid; the assay can typically be completed within 1 working day.
Disadvantages of PCR-based testing are the expense, the potential for false reactions and the requirement for targeting clinical specimens in which the microorganism is present. False-positive reactions often result from contamination of the reaction mixture with amplification products from previous reactions. Typically laboratories performing diagnostic PCR control this type of contamination by physical methods (McCreedy and Callaway 1993) or chemical methods (Persing and Cimino 1993). False-negative results are often the result of inadequate purification of the nucleic acids from the specimen resulting in samples containing substances that inhibit polymerase. These inhibitors limit the efficiency of the PCR, causing false-negative results. Diagnosis of infectious agents by PCR also requires careful attention to selection of animals and tissues for evaluation since the organism must be present at the time of testing and in the specimen evaluated. Improper selection of animals or specimens can result in false-negative results.
Five years ago, PCR assays existed for only a few murine pathogens. At the time of this writing, PCR-based assays are available for a number of murine pathogens. As PCR assays for all known murine pathogens become available, it is likely that they will replace conventional MAP and RAP testing because of the decreased time and expense required to obtain results. PCR-based diagnostic testing is likely to change rapidly over the next several years as newly developed detection methods and polymerases are implemented in the diagnostic setting.
Serological Detection of Infectious Agents
In the 1920s and 1930s, Landsteiner (1936) performed a series of landmark studies in which he demonstrated the remarkable specificity of the antibody response. These studies also showed the remarkable repertoire of the antibody response and the ability to use this response as an analytical tool. Serology uses this in vivo antibody response to measure and detect, in vitro, the response to antigens.
Antibody response. Infant rodents are protected against most infections to which their mothers are immune by antibody passed through the colostrum (Abimiku and Dolby 1987; Barthold and others 1988). Infants born to mothers with no colostral antibodies, such as specific pathogen-free (SPF1) animals, have no protection against most infectious agents. At weaning, the previously protected infant loses its protection (Arango-Jaramillo and others 1988) and becomes susceptible to pathogens in its environment. Infections endemic in the colony are maintained through contact with excretions or contaminated fomites from previously infected weanling populations. The infected weanling mounts an initial immunoglobulin M (IgM1) response, which generally lasts for days or weeks (Rodriguez and others 1996). This response generally becomes primarily immunoglobulin G (IgG1) after the first week or 2 and continues for months or years, particularly in the presence of continued antigenic stimulation by infectious agents in its environment (Bachmann and others 1996). The ability of the serological assay to detect this antibody response is therefore dependent on its ability to detect the newly arising IgM and or IgG response to the antigens of the infectious agent.
The antibodies secreted by the plasma cell gain in their strength of binding (avidity) and the number of antigens to which a response is mounted---and therefore the ability to detect the response---increases with time in most situations. Earlier detection is facilitated by the ability to detect IgM.
Antibody detection. Antigens are generally used to detect the immunoglobulins circulating in the blood, and the appropriateness of the respective antigen is paramount to the ability to detect specific antibodies. Various types of antigens are used in serological assays, varying from crude cellular extracts, to purified virus particles, to purified virus proteins. These antigen preparations are of value only if they can detect the immune response to the infecting agent. A cellular extract would be expected to have both structural and nonstructural antigens in addition to most of the cellular proteins or growth medium, whereas the purified agent would have primarily structural proteins along with the milieu proteins carried along during purification. Proteins prepared by recombinant technology must have the proper structure and antigenicity to detect the various strains of the infecting agent. The purified protein has the advantage of allowing a more concentrated presence in the assay and thus, theoretically, the ability to detect lower levels of antibody. This would be particularly true on the surface of a plate well used for enzyme-linked immunosorbent assay (ELISA1).
Assays. Along with the selected antigen preparations, the selected type of assay largely determines the serologist's ability to detect the antibody of interest. In laboratory animal assays, the hemagglutination inhibition (HAI1), ELISA, and immunofluorescence assay (IFA1) are commonly used. The complement fixation test is rarely used at the time of this writing because of its complicated format and general insensitivity. The HAI is based on the ability of certain viruses to agglutinate red blood cells (RBCs1) and form a lattice network that causes the RBCs to precipitate in a specific pattern. The purpose of the assay is to detect antibody against the viral hemagglutinin. If the HAI antibody is present in the mixture, the formation of the precipitin pattern is inhibited and a button of RBCs is formed on the bottom of the tube or well. This assay, which is limited to viruses with a protein capable of attaching to RBCs, is very useful in the differentiation of viral strains because of its specificity. Examples are the parvoviruses wherein Kilham rat virus and Toolan's H-1 are differentiated by HAI antibody, which is specific for the hemagglutinin of each virus. In addition, 1 means of differentiating minute virus of mice (MVM1) from mouse parvovirus (formerly known as mouse orphan parvovirus) (Smith and others 1993) is likewise based on the HAI. The assay is often insensitive and may yield false-positive results because of nonspecific inhibitors of hemagglutination. However, the HAI assay for polyoma was reported to detect 1 infectious unit of virus (Rowe and others 1959).
The ELISA (most commmonly the indirect ELISA) involves binding the antigen to a solid phase---in this case, a microtiter well---and then adding serum suspected of containing antibody to that antigen. If the antibody is present, it will bind to available antigen and excess serum proteins will be washed away. A significant cause of assay sensitivity is the enzyme used to label the second antibody (typically immunoglobulin such as anti-mouse and -rat), which converts its substrate from a colorless form to a spectrophotometrically measured colored product. Importantly, the enzyme is not used up by this reaction but continues to act upon substrate and result in amplification of the reaction. Within the limits of an optimized assay, the optical density (OD1) of the colored product is proportional to the amount of antibody in the test serum. When an enzyme conjugate is used that detects both IgM and IgG, the assay can detect the antibody response reasonably early in the infectious process. However, detection of an antibody response generally requires 1 week or more after infection for the level to be detectable and often longer because the timing of sampling is usually not based on the observation of illness.
ELISA may give erroneous results for a number of reasons. First, an antigen with impurities, which are present in virtually all antigens in use at the time of this writing, may bind nonspecific immunoglobulins arising from autoimmune responses (Hall and others 1992). Second, naturally circulating antibodies (such as those circulating against blood group material) may be present and may cross-react with the impurities or the antigen itself to give a falsely elevated OD. The binding ability of most ELISA plates is due to irradiation of the polystyrene surface, and variation in the intensity may cause excess binding in certain wells leading to nonspecific or "hot" wells. Any of these factors can make interpretation of the reaction difficult at times, especially when the specimens are coded as to origin and the tester does not know whether the positive reaction is 1 of many negatives from the same rack or room or 1 animal from a room of positives. In such instances, knowledge of the epidemiology and source of a specimen can often alert the tester to a possible false-positive reaction. Communication of this information to the testing laboratory can be very important in ensuring proper interpretation of results.
The IFA is often performed with infected and uninfected cells fixed to a multiwell slide. The test serum is added to the slide and unbound immunoglobulin is washed away. As in the ELISA, a labeled second antibody, which is added to the wells next, will bind to immunoglobulins attached to the antigens in the infected cells. In this case, however, the label is not an enzyme but is instead a fluorescent dye. When exposed to near ultraviolet light, the dye fluoresces in the visible spectrum and is observable in the microscope. One distinct advantage of the IFA is that the type of fluorescence (granular or smooth) and the location of the fluorescence (nuclear or cytoplasmic) can be observed as typical of the virus in question. Antibody directed against cell components, such as the cytoplasmic membrane, mitochondria, or DNA, can confuse the untrained and occasionally the trained eye because of similarity to the expected specific fluorescence. Furthermore, this nonspecific fluorescence can mask specific fluorescence and cause the test to be uninterpretable. Because the IFA is more labor intensive than the HAI or ELISA, it is often used as a confirmatory rather than primary assay.
Times and reasons to use serology. Serology is the most frequently chosen methodology for detecting the presence of infectious agents for several reasons. Because many infections (particularly in adult rodents) are asymptomatic, an infectious process taking place in the animal may be unknown. Thus, taking samples for testing by technologies that look for infectious virus by isolation or nucleic acid detection and that assume the presence of the agent may not be optimally timed. Institutions that cannot afford to pay the costs of surveillance by virus detection methods, which rely on the presence of the virus, may be able to afford the lower costs of serology. Additionally, collecting a relevant sample is more readily accomplished for serology than for many virus detection methods. However, in the case of active disease and disease signs, methods such as PCR may detect infection before serological techniques and may pinpoint the etiology sooner. Where antibody is not developed except after prolonged time and where serological techniques are insensitive, an assay such as PCR may be the best option. At the time of this writing, the ability to screen for a large number of agents still relies on initial serological testing to help determine what might be looked for by other methodologies.
Interpretation of a positive serological result requires looking at several aspects of the finding. The result is probably real if (1) adequate sampling has been done, (2) more than a few animals have been tested, and (3) a high frequency of positives are found. Relatively high HAI, ELISA, or IFA titers (or OD values if available) will also add credence to the result. The presence of the agent in other parts of the facility, signs in the animals suggestive of the agent, and confirmation by alternative assays all suggest that the finding is not a false-positive. Without this sort of corroborative evidence, it is probably wise not to take action until sufficient testing and corroborative evidence can be garnered to confirm the presence of an infectious agent. In some cases, a wait-and-see approach may be the only alternative if other, supporting evidence cannot be obtained. We recommend a test 2 to 3 weeks later to look for increasing incidence and titer of positives.
Most commercial laboratories offer a series of panels designed for different purposes. These profiles start from a basic one that includes the most prevalent agents one might encounter in laboratory rodents to complete screens that allow the user to survey the colony for baseline purposes. Of course, when using kits or a vendor, custom panels that fit institutional needs can be utilized. Once a screen has been carried out, tests for certain agents might be performed only on a yearly or longer basis, with more frequent testing for those agents more likely to be encountered. Testing frequency depends on tolerance for risk and available funds. If it is clear that a particular agent is endemic (present in all animals and cycling through generations) in the colony and the user can tolerate its presence, then further testing for that agent is unnecessary unless the goal is to "clean up" the situation.
TARGETED TESTING
To develop an effective and practical microbiological monitoring program, it is necessary to carefully choose the agents for which to screen, the type and number of animals to be sampled, and the sampling frequency. Program implementation is accomplished by systematically recording these choices and incorporating them into testing schedules.
Selection of Infectious Agents
We have concluded that it is impractical and of questionable value to define specific lists of microbes or to definitively classify agents with regard to virulence and research effects because in many cases, the information to make such choices or distinctions is incomplete or simply not available. Moreover, the reported disease or research effects produced by experimental or even natural infections with a microbial species may be atypical. Substantial variation in the virulence of different isolates (or strains) of various microbial species has been amply documented (Barthold and others 1976; Evans and Evans 1996; Falkow 1997; Riley and others 1994). Finally, host factors (in particular, the immune status of the host) can dramatically alter the outcome of infection. As mentioned in the Introduction, agents such as P. carinii, which have no discernable effects on normal laboratory animals, can be highly pathogenic for immunocompromised rodents. The selection of infectious agents to be excluded from rodent colonies is determined by the animal health and research effects of an agent and by the desired colony microbiological status. General categories used to describe the microbiological status of laboratory animals are gnotobiotic, SPF, and conventional (Foster 1980; Trexler 1983). The latter category is composed of animals that have an undefined or nominally defined microbial flora that includes common rodent pathogens. Conventional rodents are generally considered unfit for research and hence, will not be discussed further.
The word gnotobiotic is derived from the Greek words gnotos, meaning known, and biota, referring to the microflora. Gnotobiotic animals have a microflora that is entirely known; they may be axenic (germ free) or associated with a limited flora of nonpathogenic bacteria (Trexler 1983). Axenic rodents are not actually germ free since they inherit endogenous retroviruses such as murine leukemia viruses from their parents.
As might be imagined, keeping rodents axenic is extremely difficult. It requires their rederivation by cesarean section or embryo transfer and their maintenance in isolators supplied with sterilized food, water, and bedding. Monitoring must be extremely rigorous because contamination with any microorganism renders the rodents nonaxenic. An autochthonous gastrointestinal flora consisting of anaerobic and microaerophilic bacteria is essential for normal physiology and disease resistance. Consequently, axenic rodents are physiologically abnormal and particularly susceptible to disease (Schaedler and Orcutt 1983). For these reasons, they are rarely used in research.
Rodents are kept axenic after rederivation to simplify detection of vertically transmitted microbial contaminants. They are then associated with a limited, defined flora of nonpathogenic bacteria that are representative of the autochthonous gastrointestinal flora, such as the Charles River-altered Schaedler flora (CRASF1). Like axenics, associated gnotobiotes must be kept in isolators or microisolation cages and monitored frequently for microbial contamination.
CRASF-associated mice and rats are physiologically and immunologically more normal than axenics. The cecal enlargement typical of axenic animals is less pronounced after CRASF association. In contrast to axenics, the enhanced disease resistance of CRASF-associated rodents allows them to be transferred to barrier rooms for large-scale production. Barrier rooms have design features and practices that are intended to prevent adventitious infections. Air and water are filtered, supplies including food and bedding are disinfected, and animal care technicians shower and gown before entering the animal production area. Nonetheless, once in a barrier room, rodents kept in open cages become associated with a complex flora consisting of hundreds of bacteria and other microbes. Barrier-reared rodents are therefore NOT gnotobiotic. They are instead referred to as "SPF" to indicate they have tested negative for a limited list of exogenous viruses, bacteria, and parasites that may cause disease or otherwise interfere with research.
Although the term SPF might be applied legitimately to a colony from which only 1 pathogen has been excluded, in practice, SPF rodent colonies are free from an extensive list of agents. This list has grown dramatically in recent decades for several reasons:
- 1. Advances in microbiology and, more recently, molecular biology have facilitated the discovery and study of infectious agents and the development of more sensitive diagnostic tests.
2. Many published studies have shown that even nonpathogenic infectious agents can profoundly alter in vivo and in vitro biological responses. In fact, many infectious agents indigenous to rodents, in particular viruses, were discovered as latent contaminants of biological materials that interfered with research (Bonnard and others 1976; McKisic and others 1993; Riley and others 1960; Rowe, and others 1962). As investigators and laboratory animal professionals have become more aware of the research implications of adventitious infections, they have demanded mice and rats increasingly more free from recognized pathogens.
3. The supply of SPF rodents has increased because the practices of cesarean rederivation and isolator/barrier maintenance of colonies have been widely adopted by breeders. With the advent of microisolation cages and ventilated racks, users are better able to maintain the SPF status of their laboratory rodents.
SPF rodent colonies are expected to be free from ectoparasites, metazoan endoparasites, and pathogenic enteric protozoa. They are also expected to test negative for antibodies to most exogenous viruses, regardless of pathogenicity. This is because viruses are obligate intracellular parasites that alter the metabolism of the host cells they infect. These metabolic changes may be cytotoxic; however, even apparently benign, noncytolytic infections can alter nonvital cell functions (Oldstone and others 1982).
The bacteria that must be excluded from an SPF colony depend on the immune status of the rodents in it. As mentioned above, immunocompetent SPF rodents kept in open cages in barrier rooms develop a complex bacterial flora but remain healthy and suitable for most research provided they are kept free from infection with a small number of "primary" pathogens. Some of these, like Salmonella spp., are pathogenic for many animal species including humans (Clarke and Gyles 1993). Others, like M. pulmonis, appear to be rodent-specific pathogens (Lindsey and Cassell 1973). For immunodeficient rodents such as athymic nude and SCID mice, the list has been expanded to include opportunists likely to cause disease in these strains. S. aureus and P. aeruginosa are ubiquitous opportunists that may enter a barrier room via animal care technicians, supplies, or animal drinking water (Gyles 1993; Kloos and Lambe 1991). Because certain important opportunists are difficult to exclude from barrier rooms, immunodeficient and valuable transgenic strains are most often housed under very strict conditions in isolators or microisolators.
It is important to keep in mind that the classification of a bacterial species as a pathogen, opportunist, or commensal is problematic because pathogenicity, instead of being species characteristic, appears to be a clonal or strain-related property that results from unique combinations of virulence genes (Falkow 1997). Escherichia coli is normally considered to be a harmless commensal of the large intestine. It can, however, become highly pathogenic through the acquisition of plasmid or bacteriophage virulence genes that code for enterotoxins and invasion factors (Evans and Evans 1996). Riley and others (1994) have found that some C. piliforme isolates produce markedly more severe Tyzzer's disease than others; this difference may be attributed to a cytotoxin that is produced only by the more pathogenic isolates. Although most Citrobacter isolates are nonpathogenic, certain variants, notably C. rodentium, have been shown to cause murine colonic hyperplasia (Barthold and others 1976).
Sampling
Accurate, meaningful results require the sampling of an adequate number of appropriate animals on a sufficiently frequent basis. The animals for testing should be representative of the microbiological condition of the colony as a whole. One way to accomplish this is to choose animals at random, if possible. For infections and positive assay results that have an age- or strain-dependent distribution, random selection might not yield the highest prevalence of positive test results. It is more practical and efficient to obtain a representative sample by selecting animals of various ages (and sexes) and from different locations in an animal breeding or holding room. Where applicable (such as in rooms with multiple strains), strains selected for health monitoring should be rotated. Alternatively, sentinel animals, typically but not always of the same rodent species as that being monitored, are tested.
To be used successfully, sentinels should be housed in a manner that maximizes their exposure to microorganisms infecting the principal animals being monitored. In general, infections are transmitted most efficiently through animal contact. Fomite transmission is also effective, whereas airborne spread can be unreliable even for highly infectious viruses (Parker and Reynolds 1968; Thigpen and others 1989; Yang and others 1995). Fomite transmission via soiled bedding, the customary method for sentinel exposure, has generally been shown to be efficacious. There is reason not to rely on soiled-bedding alone, however, as it has been shown to not transmit certain infectious agents, such as cilia-associated respiratory bacillus (Cundiff and others 1995; Mastsushita and others 1989), and to transmit others, such as Sendai virus, inefficiently (Artwhohl and others 1994; Dillehay and others 1990).
Animal care systems that control the environment at the room, rack, or cage level have an effect on the way a colony or study is sampled and the potential for spread once an agent enters a room. An open caging system in a laminar flow room has strict entry and exit requirements; however, if an agent gains entry, spread occurs just as in a conventionally housed colony. For sampling purposes, one would expect a relatively rapid spread from the focus of infection, and the presence of the infecting agent would be more readily detected. Spread in a room where entry is controlled at the rack level would be more likely to limit spread to the cages enclosed in the rack system, and detection would depend on the sampling of all racks. Control at the cage level, such as microisolator cages, would be expected to have the greatest resistance to infection and infectious spread when used in conjunction with a laminar flow change hood. Obviously, room and rack level containment, in addition to filter top cages and the hood for cage changing, adds more levels of containment. Although providing a high level of containment against infection and spread, this also limits the ability to detect infection. Under these conditions, spread would be expected to take place slowly, if at all, from the initially infected cages (unless vitiated by wild rodents or personnel actions), and sampling would have to be more extensive to detect the break. In this situation, use of bedding transfer to sentinel cages would enhance detection for many agents.
In some instances, it is helpful to use sentinels of a different rodent species to monitor colonies of a second. One such occasion is when little is known about the viruses that infect a species, which is the case for gerbils (Clark 1984). Although gerbils can be screened for viral antibodies with serological assays that detect cross-reacting antibodies (and do not rely on difficult-to-find labeled antibodies to gerbil immunoglobulin), the results of these tests---particularly the validity of negative results---are very difficult to interpret. It is arguably more meaningful to do serology on sentinel mice or rats to determine whether gerbils are shedding murine viruses that could infect SPF rodent colonies. A second circumstance is when the principal species is susceptible to infection with nonrodent viruses carried by people and other species and, in response to infection, produces cross-reacting antibodies that are detected by a rodent serological assay. Sendai virus is a parainfluenzavirus. Members of this group infect the respiratory tract of humans. In addition to being susceptible to Sendai virus, guinea pigs can be infected with human parainfluenzavirus strains (Chanock 1979). This is suspected to be the reason they are often parainfluenzavirus antibody positive (CRL 1989). These antibodies sometimes cross-react with Sendai virus, making it difficult to determine the status of guinea pig colonies with respect to this agent. Fortunately, because mice are not susceptible to other parainfluenzaviruses, they can be used as sentinels to definitively determine the Sendai virus status of guinea pig colonies. Finally, a species might be chosen as a sentinel because it is more likely than the principal species to become ill after infection. As they are uniquely susceptible to Tyzzer's disease, gerbils have been used as sentinels to detect latent C. piliforme infections in other rodent species (Gibson and others 1987).
Animal selection is influenced by diagnostic methodology. For serology, the animals sampled should of course be immunocompetent and able to mount a strong serum antibody response to infection. Such a response is typical of disease-resistant inbred strains (Brownstein and others 1981) and outbred stocks (Parker and others 1978), although in 1 study, the Sendai virus seroconversion rate for outbred sentinel mice was lower than that for inbred strains (Artwhol and others 1994). Detectable levels of serum antibodies take on average 2 to 3 weeks to develop (Parker and Reynolds 1968; Peters and Collins 1983; Smith 1983). It is therefore recommended that sentinels be kept in a colony for at least 1 month and that sick animals be allowed to convalesce for several weeks before they are tested. In the case of production colonies, retired breeders are recommended because they have had ample opportunity to become infected and sufficient time to recover and seroconvert.
For pathology, bacteriology, and parasitology, it is especially important to sample animals of multiple ages since the prevalence of infection with some bacteria and parasites is age dependent. For example, enteric protozoa are readily observed in weanlings but not in older rodents. Conversely, because the mouse pinworm Aspicularis tetraptera has a long life cycle, adult pinworms are most often found in adolescent rather than weanling mice (Wescott 1982). Along with, or as an alternative to, sampling multiple age groups, the diagnosis of certain latent infections may be facilitated by testing immunodeficient or immunosuppressed animals. Immunosuppression to provoke Tyzzer's disease is used in the diagnosis of C. piliforme infections (Riley and others 1994; Waggie and others 1981).
Guidelines for determining sample sizes have been provided by using various formulas. In essence, these formulas show that as the prevalence of infection decreases, the sample size required to detect infection with a high level of confidence increases (Dubin and Zietz 1991). According to the standard binomial distribution formula (NRC 1976), assuming a prevalence of 30% in a colony of 100 or more animals, 8 to 10 animals must be tested to have a 95% probability of detecting at least 1 positive. To achieve the same level of confidence assuming a prevalence of 10%, a sample size of 25 to 30 animals is needed (Bhatt 1980). It is important to distinguish between prevalence and incidence. Prevalence is the percentage of positive animals at a point in time in a designated area. Incidence is the percentage of new positives over a period of time. The accuracy of the results obtained with these formulas is dependent on certain assumptions that are rarely fulfilled, for example, a minimum population size of 100 and random spread of infection. The application of these formulas has become more uncertain with the change from open cages to isolators and microisolation cages. This change has resulted in smaller effective population sizes and infections that spread more slowly and with a lower prevalence. Although the aforementioned formulas provide a framework for determining sample size, deciding on the number of animals to actually monitor is a practical compromise between the desire to achieve a high degree of certainty versus the availability of animals for monitoring and the cost of testing.
The standard formulas for determining sample size do not assess how often one should test for contamination and how data from 1 time point can be projected into the future. In other words, if one is 95% confident that a colony is SPF (that is, has prevalence of infection below a certain percentage) on the day when negative test results are obtained, how confident should one be in 1 week or 1 month that the colony is still SPF? Historical data from a commercial breeder were used to calculate contamination rates for barrier rooms and isolators. It was found that contamination rates were constant (that is, they were unaffected by the length of time a room or isolator had been in use) and so could be used to predict the likelihood of infection over time. Based on historical data, the upper 99% bound for the viral contamination rate of mouse barrier rooms was estimated to be 0.17% per wk; this meant that the probability of an SPF mouse colony becoming virally contaminated after 4 wk was less than 1%. By comparison, the Staphylococcus spp. contamination rate for isolators was estimated to be 2% per wk; thus, the probability of an isolator becoming contaminated with Staphylococcus spp. after 4 wk was 8%. In summary, the frequency of testing should be adjusted according to the historical contamination rate with infectious agents that unacceptably change the colony health status (Selwyn and Shek 1994).
The chance of an unacceptable contamination depends on the husbandry method, microbiological category, agent type, and immune status of the host (Figure 2). Contamination of isolators and microisolation cages with extraneous bacteria and fungi through physical defects or inadequately disinfected supplies is more common than are adventitious viral infections. Therefore, in the case of gnotobiotic colonies in which any microbial contamination is significant, bacteriology should be performed more frequently than serology and parasitology. Bacteriology should also be performed often on SPF immunodeficient rodents for which many opportunistic bacteria are pathogenic but can be done less regularly on immunocompetent SPF colonies because most bacterial contaminants of isolators, microisolators, and barrier-reared rodents are not primary pathogens. In barrier rooms where adventitious infections are most commonly caused by viruses, serology should be performed more often than bacteriology and parasitology.
Implementing a Microbiological Assessment Program
The text below describes ways to incorporate agent selection and sampling into a microbiological assessment program. First, profiles (panels) are defined to screen primarily for those agents to be excluded from a gnotobiotic or SPF colony. A profile may be a list of related assays, organisms, or tissues. Serology profiles, for instance, consist of antibody assays identified by method and agent. Bacteriology profiles are composed of sampling sites and lists of the primary pathogens and opportunists to be found at these sites. Pathology profiles specify the tissues and organs to be examined. Profiles are typically species specific since disease processes and pathogenic infectious agents vary by species. Several profiles may be defined to reflect the frequency with which certain infectious agents have been found. In the case of serology, basic profiles that include commonly found viruses are performed more often than comprehensive profiles to which rarely detected agents have been added. Second, test protocols are constructed by combining profiles with sample information, including number, age, and strain (Figure 3). Finally, testing frequencies are combined with test protocols to form schedule templates (Figure 4). Schedule templates are assigned to colonies to create schedules, and this may be done manually or by computer (Figure 5). On the dates indicated in the schedule, the number and type of samples designated in the protocol are collected and submitted for the test profiles specified in the protocol to a diagnostic laboratory. A submission form containing the protocol information should be sent with the samples. Result reports should be analyzed and filed in an organized fashion (for example, according to facility, room, and species). Ways to interpret results and actions prompted by unexpected positive results are reviewed below.
INTERPRETATION
When starting out with gnotobiotic or SPF rodents, the interpretation of diagnostic test results is mostly qualitative. The goal is to determine whether the animals tested have or have not been exposed to an infectious agent. Accurate quantification of antibody levels or numbers of bacteria, for example, is important only to the extent that clearly negative and positive results are easier to interpret than equivocal results near the dividing line between positive and negative.
The ideal test is one that in all cases clearly distinguishes between exposed and unaffected animals. To our knowledge, there is no such assay. Typically, a certain percentage of results are inaccurate in that samples from unaffected animals may give false-positive reactions whereas those from exposed animals may yield false-negative results (Figure 6; Tyler and Cullor 1989).
A sensitive assay is one that produces a low percentage of false-negative results or, conversely, a high percentage of true-positive results in tests performed on exposed animals. A specific assay is one that gives a low percentage of false-positive results or, conversely, a high percentage of true-negative results in tests performed on unaffected animals (Figure 7; see also Zweig and Robertson 1987).
Besides being a consequence of the limits of test sensitivity and specificity, false-positive and false-negative results can be due to sample selection and laboratory errors. Examples of sample selection errors are shown in Table 1. Myriad laboratory errors can cause inaccurate results including improper sample preparation and storage, sample mix-ups, deviation from accepted procedures, and transcription mistakes during report preparation.
It is especially important to confirm unexpected positive findings before taking any action. This is accomplished by repeat testing of the positive samples, testing additional samples, and using alternative assays and diagnostic methodologies to corroborate primary test results. For example, sera that are M. pulmonis ELISA positive might be repeat tested for specific antibodies by IFA. Additional animals from the suspect colony could be cultured for mycoplasma and examined grossly and microscopically for lung lesions. Finally, mycoplasma isolates could be identified as M. pulmonis with species-specific antisera.
Positive results are questionable when they are borderline or when the prevalence of positive animals is low. Borderline reactions near the positive-negative dividing line are those most likely to be inaccurate (Figure 6). The predictive value of a positive result (PV+) is the percentage of all positive results that are true positive. As the prevalence of positive results decreases, so does PV+. In the example shown in Figure 8 (see also Zweig and Robertson 1987), PV+ for an assay with a specificity of 95% is calculated when the prevalence of positives is 10% and when it is 1%. When the prevalence is 10%, PV+ is 68.5%. In other words, one would expect positive results to be incorrect 31.5% of the time. With a prevalence of 1%, only 16.5% of the positive results are expected to be real. In general, when the prevalence is less than 15%, PV+ is too low to be meaningful (La Regina and Lonigro 1988).
Once results are confirmed, the next step is to decide whether to keep the affected colony. Of course, animals infected with a zoonotic agent should be euthanized, decontaminated, and then safely discarded. Because pathogens often suppress the immune response and immunosuppression can persist even in recovered animals, the use of infected rodents in intricate immunological research should be avoided. It is clearly contraindicated to do research involving tissues or organs that are the targets of an adventitious infection. Rodents that have undergone an adventitious viral infection should not be used for passaging cell lines or as sources of tissues and fluids for subsequent experiments. A virus might contaminate these materials, especially if it causes a persistent infection, has a broad host range, or has a predilection for replicating in rapidly dividing tumor cells (Collins and Parker 1972).
Nevertheless, one famous virologist/geneticist accepted the fact that several agents were infecting his colony. He knew that these endemic infections did not affect the health of his animals and did not play a role in the genetic inheritance studies that he did. He also knew that because of the circumstances under which he had to work, it would be very difficult to prevent the recurrence of infections, so he decided to accept the status quo. However, for many studies, infectious agents can and do affect the ability to obtain valid research results. For this reason, accreditation and granting agencies prefer research work to be done in host animals free from infectious variables that could call such studies in question, after the expenditure of large sums of money and time.
If an animal population has become infected, an animal care manager should not start a new SPF colony before investigating the source of the infection and making the necessary procedural and facility modifications to prevent a recurrence. Possible sources include the breeder colony, exposure during shipment, wild rodents, and contaminated biological materials. Examples of practices that should be implemented, if they are not already in place, are pest control and the testing of biological materials to be injected into rodents for extraneous murine viruses by the MAP test or by PCR. Isolators and microisolators offer a high degree of protection against adventitious infections provided that the appropriate procedures are followed.
Some viruses, such as mouse hepatitis virus, do not persist in the recovered host and are not stable in the environment. These agents are often dealt with by waiting out the infection. In other words, breeding and importation of susceptibles are ceased and testing is done to ensure that all animals in a substantial sample have been infected. The time to ensure universal exposure can be hastened by purposely spreading the infection with bedding. This procedure is recommended only if the animals cannot be replaced, there is no health problem, and the sequelae of infection (Barthold 1986a,b) can be tolerated. At that time, approximately 4 weeks after all animals have seroconverted, cage or bedding transfer sentinels should be introduced and a thorough disinfection of the room, racks, for example, is carried out to ensure the absence of infectious virus before renewal of importation and breeding. A time-efficient alternative to suspending breeding is to start a new colony with seropositive noncontagious breeders (Brammer and others 1993).
In the case of viruses such as parvoviruses, the time to disappearance of potentially infectious virus is difficult to determine, and this procedure is not recommended. In the case of chronic agents, such as cytomegalovirus or mouse thymic virus, the effort would be fruitless. When selecting animals for testing in a disease outbreak situation in which signs of illness are present, it is best to test animals that have survived infection or do not show signs of infection. These animals will more likely have developed a detectable antibody response, and the animal showing signs will be better candidates for PCR or virus isolation.
In summary, no diagnostic test always gives accurate results. False-positive and false-negative results occur because tests are not completely specific and sensitive and because sample selection and laboratory errors occur. Consequently, it is prudent to always confirm unexpected positive findings before deciding on a course of action. This is accomplished by repeat testing of the same and additional samples using a variety of methodologies. Once positive results have been confirmed, the decision to keep or depopulate the colony should be based on the public health significance and research effects of the infectious agent and the ability to contain or prevent a recurrence of the infection within the facility.
NEWLY RECOGNIZED INFECTIOUS AGENTS
During the 1990s, several previously undescribed microorganisms have been identified in laboratory mice and rats. These include newly recognized and identified mouse and rat parvovirus species and several members of the genus Helicobacter. Although these organisms have been only recently identified, review of retrospective data indicates that these organisms were present in laboratory rodents for many years. Thus, instead of newly emerging infectious agents in laboratory rodent populations, these organisms more accurately represent organisms that have been present, but undetected, in mouse and rat populations.
Parvoviruses
Newly recognized parvovirus species have been identified in mice and rats. These viruses are distinct from the previously described and characterized rodent parvoviruses MVM, H-1, and RV. Evidence of the existence of undescribed rodent parvoviruses first emerged in the 1980s, when several diagnostic testing laboratories began to report atypical serological results. Serum reacted in ELISA or IFA with antigens from mouse and rat parvoviruses; however, HAI specific for MVM, H-1, and RV were negative, suggesting the presence of distinct parvoviruses. The newly recognized mouse-origin species has been designated mouse parvovirus or MPV1 (McKisic and others 1993); the rat-origin species has been designated rat parvovirus or RPV1 (Jacoby and others 1996).
Although much remains to be learned about these newly identified rodent parvoviruses, we know that MPV and RPV are distinct from MVM, H-1, and RV in their genomic sequence, pathogenesis, and potential to interfere with research (Ball-Goodrich and Johnson 1994; Besselsen and others 1996). Although natural infections with either MVM or MPV are asymptomatic, experimental infection studies suggest that MPV is less pathogenic than MVM since even neonatal and immunocompromised mice experimentally infected with MPV show no clinical signs, whereas MVM infections of neonatal mice may be lethal (Brownstein DG and others 1991; Smith and others 1993). Another difference between MPV and MVM is the apparent persistence of MPV. Infections with MPV in adult mice persist for at least 9 weeks after infection, whereas infections with MVM (Jacoby and others 1995; Smith and others 1993) are short lived (Smith and Paturzo 1988). MPV is also known to modulate in vitro and in vivo immune responses (McKisic and others 1993, 1996). In contrast, MVM is documented to modulate only the in vitro immune response (Bonnard and others 1976). The persistence of MPV and its ability to modulate in vivo immune responses suggest that MPV's potential to interfere with research may be greater than MVM.
RPV appears to be more closely related to MPV than to the other rat parvovirus species, H-1 or RV. However, even less is known about the pathogenesis of RPV infections. In almost all cases, natural infections with any of the rat parvovirus species cause no clinical or histological evidence of disease. The one exception is RV infections, which have manifested in clinical disease. Interestingly, it appears that multiple subspecies of RPV exist whose genomic sequences differ considerably (L. K. Riley, personal communication, 1998); however, it is not known whether the pathogenesis or virulence of RPV strains differ.
Diagnosis of MPV infections in mice relies on serology or PCR assays since no clinical or histological disease is evident. Serology can be accomplished by ELISA or IFA using either MPV, Kilham rat virus (KRV1), or MVM as the antigen. Alternatively, an ELISA based on recombinant nonstructural protein (NS1) can be used (Riley and others 1996). Because of the cross-reactivity among rodent parvovirus species, ELISA and IFA assay results may not delineate the species of the parvovirus. A MPV-specific HAI is used as a secondary test. Positive HAI results indicate that MPV is the parvovirus species involved in the infection. Similarly, diagnosis of RPV infections is accomplished by testing serum samples by ELISA or IFA with KRV, H-1,or recombinant NS1 as the antigen. Samples positive for ELISA or IFA are retested with HAIs for KRV and H-1. Negative HAI results indicate that the parvovirus involved is not KRV or H-1 and suggest exposure to RPV. Development of a RPV-specific HAI is needed but has been hampered by difficulties associated with in vitro propagation of sufficient amounts of the virus. PCR assays have also been developed to aid in diagnosis of rodent parvovirus infections. A variety of parvovirus PCR assays are available. A universal rodent parvovirus assay detects any of the parvovirus species but does not delineate the specific virus species involved. Species-specific assays are also available for MVM, MPV, KRV, H-1, and RPV, which identify individual parvovirus species (Besselsen 1995a,b).
Helicobacters
Since 1994, a number of Helicobacter species have been identified in rodents. To date, 2 species, Helicobacter hepaticus and Helicobacter bilis, have been documented to cause disease syndromes including chronic active hepatitis, hepatic neoplasia, inflammatory bowel disease, proliferative colitis, and rectal prolapse (Fox and others 1995; Ward and others 1994a,b 1996). Additional, as yet unnamed, Helicobacter species have recently been identified that appear to be pathogenic in immunodeficient mice, causing lesions similar to those described for H. hepaticus (unpublished data). Several Helicobacter species also exist that, until the time of this writing, have not been associated with disease, raising the possibility that some members of this genus are commensal organisms, which do not cause disease (Mendes and others 1996; Schauer and others 1993; Shen and others 1997).
H. hepaticus has been documented to cause hepatic disease in immunocompetent and immunocompromised mice. Histological lesions vary from foci of hepatocellular necrosis with or without associated infiltration of inflammatory leukocytes to extensive oval cell hyperplasia with occasional bile ductule formation. H. hepaticus has also been associated with intestinal disease including a high incidence of rectal prolapse and chronic proliferative typhlocolitis and proctitis in immunodeficient mice (Ward and others 1996).
Much less is known about H. bilis, although it has been shown to colonize the intestinal tract and the liver of several strains of mice (Fox and others 1995; Franklin and others forthcoming). Hepatic colonization has been associated with mild hepatitis. Like H. hepaticus, H. bilis has also been linked to intestinal disease with lesions virtually identical to those found in H. hepaticus-infected SCID mice (Franklin and others forthcoming).
Intestinal colonization with H. hepaticus occurs in a wide range of mouse strains/stocks and does not appear to be gender or age specific. However, susceptibility to hepatic disease associated with H. hepaticus varies depending on the strain, stock, gender, and age of the mouse. Immunodeficient strains appear to be particularly susceptible to development of intestinal lesions (Ward and others 1996). H. hepaticus causes hepatic lesions more frequently in male mice, and the lesions are typically more severe in males than females (Fox and others 1996; Ward and others 1994a,b). Since lesions progress with time, more severe lesions are often found in older mice.
Less is known about the host range of H. bilis. Like mice infected with H. hepaticus, only a small percentage of intestinally colonized mice exhibit hepatic lesions. When present, hepatic lesions occur predominately in older animals. At the time of this writing, a gender bias in susceptibility to hepatic lesion development associated with H. bilis infection has not been established nor has the susceptibility of specific mouse strains or stocks been examined.
The effects these agents may have on research investigations that utilize infected mice are largely unknown at the time of this writing. The major concern with the use of H. hepaticus- and H. bilis-infected mice is in studies involving immunodeficient mice and in long-term studies involving old mice. Both of these groups have demonstrated a much higher frequency of intestinal and hepatic lesions as well as an increased incidence of hepatocellular tumors, which may confound interpretation of experimental results.
Several methods have been employed to detect Helicobacter infections in laboratory animals, including histopathology, culture, PCR assays, and serology. Histopathology using a modified Steiner's silver stain can identify spiral organisms in tissues. However, it is difficult by histopathology to definitively identify Helicobacters in the intestine among the myriad of normal flora bacteria present in the gastrointestinal tract of rodents. Because Helicobacter organisms can be found in the liver in less than 10% of infected mice, histopathology lacks sensitivity for routine diagnosis of Helicobacter infections in rodents. Bacterial culture provides a sensitive method for detection of Helicobacters from liver, intestine, and fecal specimens. The major drawback of culture as a diagnostic method is the fastidious and slow-growing nature of the organisms, which require microaerophilic conditions and 5 to 7 days for primary isolation. Diagnosis of Helicobacter infections by culture is further complicated by the presence of normal flora, since the preferred site for isolation of the bacteria is the gastrointestinal tract. PCR assays have been developed by several laboratories as an alternative to culture for detection of Helicobacters in liver, intestinal, and fecal samples (Battles and others 1995; Beckwith and others 1997; Riley and others 1996; Shames and others 1995). These assays have the advantage of being as sensitive as culture but more specific and rapid. A serological assay has recently has been developed for detection of H. hepaticus infection in mice (Livingston and others 1996). The obvious advantage to serological testing for H. hepaticus is the rapid turnaround and decreased expense of testing compared with histopathology, culture, and PCR. Although these methods can be used to detect Helicobacter infections in laboratory mice, personnel should be aware of the sensitivity and specificity of the diagnostic method used to ensure that decisions regarding animal care and use of the animal in specific protocols will not be compromised.
At the time of this writing, murine Helicobacters are believed to be spread by fecal-oral transmission. The organism appears to be readily transmitted since greater than 95% of the mice in an enzootically infected colony are colonized. Several antibiotic treatment regimens have been explored to eliminate Helicobacter from mouse colonies (Foltz and others 1996; Foltz and others 1995; Russell and others 1995).
Additional Emerging and Newly Recognized Agents
It is likely that additional microorganisms and pathogens presently unknown in laboratory rodents will be identified in the future. Three factors may play a role in identification and recognition of new infectious agents. First, recent advances in diagnostic methodologies have enhanced our ability to detect previously unrecognized microorganisms. For example, with the advent of PCR technology, microbiologists have identified a myriad of organisms in the environment, in animals, and in people who were not known previously because they could not be cultured. PCR now permits detection and identification of these uncultivable organisms. As the use of PCR technology in laboratory animal medicine increases, it is likely that uncultivable infectious agents or agents that we have not yet learned to culture will be identified in laboratory mice and rats.
Second, the explosion in the development of transgenic and knockout mouse strains may also be an important factor in identification of new infectious agents. The immune system of these genetically altered mice is often aberrant. Organisms that are held in check by the immune system in normal mice may proliferate and cause disease in transgenic and knockout mice. Thus, previously undescribed organisms and diseases may be identified in the future.
Finally, because of the exposure of laboratory mice and rats to humans (either investigators or veterinary care staff), the possibility also exists that organisms of human origin may be transmitted to laboratory animals, providing yet another source of new rodent infections.
1Abbreviations used in this paper: CRASF, Charles River-altered Schaedler flora; ELISA, enzyme-linked immunosorbent assay; HAI, hemagglutination inhibition; IFA, immunofluorescence assay; IgG, immunoglobulin G; IgM, immunoglobulin M; KRV, Kilham's rat virus; MAP, mouse antibody production; MPV, mouse parvovirus; MVM, minute virus of mice; OD, optical density; PCR, polymerase chain reaction; RAP, rat antibody production; RBC, red blood cell; RPV, rat parvovirus; SCID, severe combined immunodeficiency; SPF, specific pathogen free.
REFERENCES
Abimiku AG, Dolby JM. 1987. The mechanism of protection of infant mice from intestinal colonisation with Campylobacter jejuni. J Med Microbiol 23:339-344.
Arango-Jaramillo S, Wisseman CL Jr., Azad AF. 1988. Newborn rats in the murine typhus enzootic infection cycle: Studies on transplacental infection and passively acquired maternal antirickettsial antibodies. Am J Trop Med Hyg 39:391-397.
Armstrong My, Smith AL, Richards FF. 1991. Pneumocystis carinii pneumonia in the rat model. J Protozool 38:136S-138S.
Artwohl JE, Cera LM, Wright MF. Medina LV, Kim LJ. 1994. The efficacy of dirty bedding sentinel system for detecting Sendai virus infection in mice: A comparison of clinical signs and seroconversion. Lab Anim Sci 44:73-75.
Bachmann MF, Odermatt B, Hengartner H, Zinkernagel RM. 1996. Induction of long-lived germinal centers associated with persisting antigen after viral infection. J Exp Med 183:2259-2269.
Ball-Goodrich LJ, Johnson E. 1994. Molecular characterization of a newly recognized mouse parvovirus. J Virol 66:6476-6486.
Barthold SW. 1986a. Research complications and state of knowledge of rodent coronaviruses. In: Hamm Jr TE, editor. Complications of Viral and Mycoplasma Infections in Rodents to Toxicology Research and Testing. Washington DC: Hemisphere Press. p 53-89.
Barthold SW. 1986b. Mouse hepatitis virus biology and epizootiology. In Bhatt PN, Jacoby RO, Morse HC III, New AE, editors. Viral and Mycoplasmal Infections of Laboratory Rodents: Effects on Biomedical Research. Orlando FL: Academic Press. p 585-588.
Barthold SW, Beck DS, Smith AL. 1988. Mouse hepatitis virus and host determinants of vertical transmission and maternally-derived passive immunity in mice. Arch Virol 100:171-183.
Barthold SW, Coleman GL, Bhatt P, Osbaldiston GW, Jonas AM. 1976. The etiology of transmissible murine colonic hyperplasia. Lab Anim Sci 26:889-894.
Battles JK, Williamson JC, Pike KM, Gorelick PL, Ward JM, Gonda MA. 1995. Diagnostic assay for Helicobacter hepaticus based on nucleotide sequence of its 16S rRNA gene. J Clin Microbiol 33:1344-1347.
Beckwith CS, Franklin CL, Hook RR, Jr, Besch-Williford CL, Riley LK. 1997. Fecal PCR assay for diagnosis of Helicobacter infections in laboratory rodents. J Clin Microbiol 35:1620-1623.
Besselsen DG, Besch-Williford CL, Pintel DJ, Franklin CL, Hook RR, Jr., Riley LK. 1995a. Detection of H-1 parvovirus and Kilham rat virus by PCR. J Clin Microbiol 33:1699-1703.
Besselsen DG, Besch-Williford CL, Pintel DJ, Franklin CL, Hook RR, Jr., Riley LK. 1995b. Detection of newly recognized rodent parvoviruses by PCR. J Clin Microbiol 33:2859-2863.
Besselsen DG, Pintel DJ, Purdy GA, Besch-Williford CL, Franklin CL, Hook RR, Jr., Riley LK. 1996. Molecular characterization of newly recognized rodent parvoviruses. J Gen Virol 77:899-911.
Bhatt P. 1980. Virus infections of laboratory rodents. Lab Anim 9:43-47.
Bhatt PN, Jacoby RO. 1985. Epizootiological observations of natural and experimental infection with sialodacryoadenitis virus in rats. Lab Anim Sci 12:196-209.
Bonnard GD, Manders EK, Campbell DA, Herbermann RB, Collins MJ. 1976. Immunosuppressive activity of a subline of the mouse EL-4 lymphoma: Evidence for a minute virus of mice causing the inhibition. J Exp Med 143:187-205.
Bonnard CD, Manders EK, Campbell DA, Herberman RB, Collins MJ. 1976. Immunosuppressive activity of a subline of the mouse EL-4 lymphoma. J Exp Med 143:187-205.
Brammer DW, Dysko RC, Spilman SC, Oskar PA. 1993. Elimination of sialodacryoadenitis virus from a rat production colony by using seropositive breeding animals. Lab Anim Sci 43:633-634.
Brownstein DG, Smith AL, Jacoby RO, Johnson EA, Hansen G, Tattersall P. 1991. Pathogenesis of infection with a virulent allotropic variant of minute virus of mice and regulation by host genotype. Lab Invest 65:357-364.
Brownstein DG, Smith AL, Johnson EA. 1981. Sendai virus infection in genetically resistant and susceptible mice. Am J Pathol 105:156-163.
Chanock RM. 1979. Parainfluenza viruses. In: Lennette EH, Schmidt NJ, editors. Diagnostic Procedures for Viral, Rickettsial and Chlamydial Infections. 5th Ed. Washington DC: American Public Health Association. p 611-632.
Clark JD. 1984. Biology and diseases of other rodents. In: Fox JG, Cohen BJ, Loew FM, editors. Laboratory Animal Medicine. Orlando, FL: Academic Press. p 183-206.
Clarke RC, Gyles CL. 1993. Salmonella. In: Gyles CL, Thoen CO editors. Pathogenesis of Bacterial Infections in Animals. Ames IA: Iowa State University. p 133-153.
Collins MJ, Parker JC. 1972. Murine virus contaminants of leukemia viruses and transplantable tumors. J Natl Cancer Inst 49:1139-1143.
CRL [Charles River Laboratories]. 1989. Revised Reporting on Simian Virus 5 (SV5) and Sendai Virus in Guinea Pigs. Reference Paper, Spring.
Cundiff DD, Riley LK, Franklin CL, Hook RR, Besch-Williford C. 1995. Failure of a soiled bedding sentinel system to detect cilia-associated respiratory bacillus infection in rats. Lab Anim Sci 45:219-221.
de Souza M, Smith AL. 1989. Comparison of isolation in cell culture with conventional and modified mouse antibody production tests for detection of murine viruses. J Clin Microbiol 27:185-187.
Dillehay DL, Lehner NDM, Huerkamp MJ. 1990. The effectiveness of microisolator cage system and sentinel mice for controlling and detecting MHV and Sendai virus infections.
Dubin S, Zietz S. 1991. Sample size for animal health surveillance. Lab Anim 20:29-33.
Evans DJ, Evans DG. 1996. Escherichia Coli in diarrheal disease. In: Baron S, editor. Medical Microbiology. Galveston TX: University of Texas Medical Branch at Galveston. p 337-345.
Falkow S. 1997. What Is a Pathogen? ASM News 63(7):359-365.
Foltz CJ, Fox JG, Yan L, Shames B. 1995. Evaluation of antibiotic therapies for eradication of Helicobacter hepaticus. Antimicrob Agents Chemother 39:1292-1294.
Foltz CJ, Fox JG, Yan L, Shames B. 1996. Evaluation of various oral antimicrobial formulations for eradication of Helicobacter hepaticus. Lab Anim Sci 46:193-197.
Foster HL. 1980. Gnotobiology. In: Baker HJ, Lindsey JR, Weisbroth SH editors. The Laboratory Rat. Vol II. Research Applications, San Diego CA: Academic Press. p 43-58.
Fox JG 1977. Clinical assessment of laboratory rodents on long term bioassay studies. J Environ Pathol Toxicol 1:199-226.
Fox JG, Yan LL, Dewhirst FE,. Paster BJ, Shamers B, Murphy JC, Hayward A, Belcher JC, Mendes EN. 1995. Helicobacter bilis sp. nov., a novel Helicobacter species isolated from bile, livers, and intestines of aged, inbred mice. J Clin Microbiol 33:445-454.
Fox JG, Yan L, Shames B, Campbell J, Murphy JC, Li X. 1996. Persistent hepatitis and enterocolitis in germfree mice infected with Helicobacter hepaticus. Infect Immun 64:3673-3681.
Franklin CL, Riley LK, Livingston RS, Beckwith CS, Besch-Williford CL, Hook RR Jr. 1998. Enterohepatic Lesions in SCID Mice Infected with Helicobacter bilis. City State: Publisher. Lab Anim Sci (Forthcoming).
Fries AA. 1979. Studies on Tyzzer's disease: Acquired immunity against infection and activation of infection by immunosuppressive treatment. Lab Anim 13:143-147.
Gibson SV, Waggie KS, Wagner JE, Ganaway JR. 1987. Diagnosis of subclinical Bacillus piliformis infection in a barrier-maintained mouse production colony. Lab Anim Sci 37:786-788.
Gyles CL. 1993. Pseudomonas and Moraxella. In: Gyles CL, Thoen CO editors. Pathogenesis of Bacterial Infections in Animals. Ames IA: Iowa State University. p 248-261.
Hall WC, Ganaway JR, Rao GN, Peters RL, Allen AM, Luczak JW, Sandberg EM, Quigley BH. 1992. Histopathologic observations in weanling B6C3F1 mice and F344/N rats and their adult parental strains. Toxicol Pathol 20:146-154.
Jacoby RO, Ball-Goodrich LJ, Besselsen DG. McKisic MD, Riley LK, Smith AL. 1996. Rodent parvovirus infections. Lab Anim Sci 46:370-380.
Jacoby RO, Barthold SW. 1980. Quality assurance for rodents used in toxicological research and testing. In: Gralla EJ, editor. Scientific Considerations in Monitoring and Evaluating Toxicological Research. Washington DC: Hemisphere Pull Corp. p 27-55.
Jacoby RO, Johnson EA, Ball-Goodrich LJ, Smitt AL, McKisic MD. 1995. Characterization of mouse parvovirus infection by in situ hybridization. J Virol 69:3915-3919.
Jacoby RO, Johnson EA, Paturzo FX, Gaertner DJ, Brandsma JL, Smitth AL. 1991. Persistent rat parvovirus infection in individually housed rats. Arch Virol 117:193-205.
Jacoby RO, Lindsey JR. 1997. Health care for research animals is essential and affordable. FASEB J 11:609-614.
Kitada K, Serikawa T. 1994. Pneumocystis carinii. In: Waggie K, Kagiyama N, Ailen AM, Nomura J, editors. Manual of Microbiologic Monitoring of Laboratory Animals. 2nd ed. Bethesda MD: US Dept Health and Human Services (NIH publication no. 94-2498). p 205-208.
Kloos WE, Lambe DW. Staphylococcus. 1991. In: Balows A, Hausler WJ, Herrmann JL, Isenberg HD, Shadomay HJ, editors. Manual of Clinical Microbiology. 5th Ed. Washington DC: American Society for Microbiology. p 222-237.
Kraft V, Deeny AA, Blanchet HM, Boot R, Hansen AK, Hem A, von Herck H, Kunstyr I, Milite G, Needham JR, Nicklas W, Perrot A, Rehbinder C, Richard Y, DeVroey G. 1994. Report of the FELASA Working Group on Animal Health: Recommendation for the health monitoring of mouse, rat, hamster, guinea pig and rabbit breeding colonies. Lab Anim 28:1-12.
La Regina MC, Lonigro J. 1988. Serologic screening for murine pathogens: basic concepts and guidelines. Lab Animal 89:40-47.
Landsteiner K. 1936. The specificity of Serological Reactions. New York: Dover Publications Inc.
Lindsey JR, Cassell GH. 1973. Experimental Mycoplasma pulmonis infection in pathogen-free mice. Am J Pathol 721:63-83.
Livingston RS, Riley LK, Steffen EK, Besch-Williford CL, Hook RR Jr, Franklin CL. 1996. Serodiagnosis of Helicobacter hepaticus in mice by an enzyme-linked immunosorbent assay. ContempTop (Abstract) 35:62.
Loew FM, Fox JG. 1983. Animal health surveillance and health delivery systems. In: Foster AL, Small D, Fox JG, editors. The Mouse in Biomedical Research. Vol. III. New York: Academic Press. p 69-82.
Mastsushita S, Joshima H, Matsumoto T, Fukutsu K. 1989. Transmission experiments of cilia-associated respiratory bacillus in mice, rabbits and guinea pigs. Lab Anim 23:96-102.
McCreedy BJ, Callaway TH. 1993. Laboratory design and flow. In: Persing DH, Smith TF, Tenover FC, White TJ, editors. Diagnostic Molecular Microbiology: Principals and Applications. Washington DC: American Society for Microbiology. p 149-159.
McKisic MD, Lancki DW, Otto G, Padrid P, Snook S, Cronin DC II, Lohmer PD, Wong T, Fitch F. Identification and propagation of a putative immunosuppressive orphan parvovirus in cloned T cells. J Immunol 150:419-428.
McKisic MD, Paturzo FX, Smith AL. 1996. Mouse parvovirus infection potentiates rejection of tumor allografts and modulates T cell effector functions. Transplantation 61:292-299.
Mendes EN, Queiroz DMM, Dewhirst FE, Paster BJ, Moura SB, Fox JG. 1996. Helicobacter trogontum sp. Nov., isolated from the rat intestine. Int J System Bacteriol 46:916-921.
Morse SS. 1990. Comparative sensitivity of infectivity assay and mouse antibody production (MAP) test for detection of mouse thymic virus (MTLV). J Virol Methods 28:15-23.
Nakayama H, Oguihara S. Osaki K, Toriumi W, Fujiwara K. 1984. Effect of cyclophosphamide on Tyzzer's disease of mice. Jpn J Vet Sci 47:81-88.
Nicklas W, Kraft V, Meyer B. 1993. Contamination of transplantable tumors, cell lines, and monoclonal antibodies with rodent viruses. Lab Anim Sci 43:296-300.
NRC [National Research Council]. 1976. Long-term holding of laboratory rodents. ILAR News 19:L1-L25.
NRC [National Research Council]. 1991. Health surveillance programs. In: Infectious Diseases of Mice and Rats. Washington DC: National Academy Press. p 21-27.
NRC [National Research Council]. 1996. Guide for the Care and Use of Laboratory Animals. 7th ed. Washington DC: National Academy Press. p 59-60.
Oldstone MBA, Sinha YN, Blount P, Tishon A, Rodriguez M, Von Wedel R, Lampert PW. 1982. Virus-induced alterations in homeostasis: Alterations in differentiated functions of infected cells in vivo . Science 218:1125-1127.
Osborn JE. 1982. Cytomegalovirus and other herpesviruses. In: Foster HL, Small JD, Fox JG, editors. The Mouse in Biomedical Research. Vol II: Diseases. New York: Academic Press. p 267-293.
Parker JC, Reynolds RK. 1968. Natural history of Sendai virus infection in mice. Am J Epidemiol 88:112-125.
Parker JC, Whiteman MD, Richter CB. 1978. Susceptibility of inbred and outbred mouse strains to Sendai virus and prevalence of infections in laboratory rodents. Infect Immun 19:123-130.
Persing DH, Cimino GD. 1993. Amplification product inactivation methods. In: Persing DH, Smith TF, Tenover FC, White TJ, editors. Diagnostic Molecular Microbiology: Principals and Applications. Washington DC: American Society for Microbiology. p 105-112.
Peters RL, Collins MJ. 1983. The enzyme-linked immunosorbent assay for mouse hepatitis and rat coronaviruses. Lab Animal 12:18-25.
Quigley BH. 1992. Histopathologic observations in weanling B6C3F1 mice and F344/N rats and their adult parental strains. Tox Pathol 20:146-154.
Rehbinder C, Baneux P, Forbes D, Van Herck H, Nicklas W, Rugaya Z, Winkler G. 1996. Report of FELASA Working Group on Animal Health: Recommendations for the health monitoring of mouse, rat, hamster, gerbil, guinea pig and rabbit experimental units. Lab Anim 30:193-208.
Riley LK, Franklin CL, Hook RR, Besch-Williford C. 1994. Tyzzer's disease: An update of current information. Charles River Laboratories Technical Bulletin.
Riley LK, Franklin CL, Hook RR Jr, Besch-Williford CL. 1996. Identification of murine Helicobacters by PCR and restriction enzyme analyses. J Clin Microbiol 34:942-946.
Riley LK, Knowles R, Purdy G, Salom J, Pintel D, Hook RR Jr, Franklin CL, Besch-Williford CL. 1996. Expression of recombinant parvovirus NS1 protein by a baculovirus and application to serologic testing of rodents. J Clin Microbiol 34:440-444.
Riley V, Lilly F, Huerto E, Bardell D. 1960. Transmissible agent associated with 26 types of experimental mouse neoplasms. Science 132:545-547.
Rodriguez M del C, Torres JA, Zlotnik H. 1996. Investigation of the humoral immune response in a murine model of actinomycetoma. P R Health Sci J 15:91-95.
Rowe WP, Hartley JW, Estes JD, Huebner RJ. 1959. Studies of mouse polyoma virus infection. I. Procedures for quantitation and detection of virus. J Exp Med 109:379-391.
Rowe WP, Hartley JW, Huebner RJ. 1962. Polyoma and other indigenous mouse viruses. In: Harris RJC, editor. The Problems of Laboratory Animal Disease. New York, Academic Press. p 131-142.
Russell RJ, Haines DC, Anver MR, Battles JK, Gorelick PL, Blumenauer LL, Gonda MA, Ward JM. 1995. Use of antibiotics to prevent hepatitis and typhlitis in male scid mice spontaneously infected with Helicobacter hepaticus. Lab Anim Sci 45:373-378.
Schaedler RW, Orcutt RP. 1983. Gastrointestinal microflora. In: Foster HL, Small JD, Fox JG, editors. The Mouse in Biomedical Research. Vol III. Normative Biology, Immunology, and Husbandry. San Diego CA: Academic Press. p 327-345.
Schauer DB, Ghori N, Falkow S. 1993. Isolation and characterization of "Flexispira rappini" from laboratory mice. J Clin Microbiol 31:2709-2714.
Selwyn MR, Shek WR. 1994. Sample sizes and frequency of testing for health monitoring in barrier rooms and isolators. Contemp Top 33:56-60.
Shames B, Fox JG, Dewhirst F, Yan L, Shen Z, Taylor NS. 1995. Identification of widespread Helicobacter hepaticus infection in feces in commercial mouse colonies by culture and PCR assay. J Clin Microbiol 33:2968-2972.
Shen Z, Fox JG, Dewhirst FE, Paster BJ, Foltz CJ, Yan L, Shames B, Perry L. 1997. Helicobacter rodentium sp. Nov., a urease-negative Helicobacter species from laboratory mice. Int J System Bacteriol 47:627-634.
Small JD. 1984. Rodent and lagomorph health surveillance---Quality assurance. In: Fox JG, Cohen BJ, Loew FM, editors. Laboratory Animal Medicine. New York: Academic Press. p 709-723.
Smith AL. 1983. An immunofluorescence test for detection of serum antibody to rodent coronaviruses. Lab Anim Sci 33:157-160.
Smith AL, Jacoby RO, Johnson EA, Paturzo F, Bhatt, PN. 1993. In vivo studies with an "orphan" parvovirus of mice. Lab Anim Sci 43:175-182.
Smith AL, Paturzo FX. 1988. Explant cultures for detection of minute virus of mice in infected mouse tissue. J Tissue Cult Methods 11:45-47.
Thigpen JE, Lebetkin EH, Dawes ML, Amyx HL, Caviness GF, Sawyer BA, Balckmore DE. 1989. The use of dirty bedding for detection of murine pathogens in sentinel mice. Lab Anim Sci 29:324-327.
Trexler PC. 1983. Gnotobiotics. In: Foster HL, Small JD, Fox JG, editors. The Mouse in Biomedical Research. Vol III. Normative Biology, Immunology, and Husbandry. San Diego CA: Academic Press. p 1-17.
Tyler JW, Cullor JS. 1989. Titers, test, and truisms: Rational interpretation of diagnostic serologic testing. JAVMA 194:1550-1558.
Waggie KS, Hansen CT, Ganaway JR, Spencer TS. 1981. A study of mouse strain susceptibility to Bacillus piliformis (Tyzzer's disease): The association of B-cell function and resistance. Lab Anim Sci 31:139-142.
Waggie K, Kagiyama N, Allen AM and Nomura T. 1994. Manual of Microbiologic Monitoring of Laboratory Animals. 2nd ed. US Department of Health and Human Services Bethesda MD: National Institutes of Health (NIH publication no. 94-2498). p 1-3.
Wagner JE, Besch-Williford CL, Steffer EK. 1991. Health Surveillance of Laboratory Rodents. Lab Anim 20:40-45.
Ward JM, Anver MR, Haines DC, Benveniste RE. 1994a. Chronic active hepatitis in mice caused by Helicobacter hepaticus. Am J Pathol 145:959-968.
Ward JM, Anver MR, Haines DC, Melhorn JM, Gorelick P, Yan L, Fox JG. 1996. Inflammatory large bowl disease in immunodeficient mice naturally infected with Helicobacter hepaticus. Lab Anim Sci 46:15-20.
Ward JM, Fox JG, Anver MR, Haines DC, George CV, Collins MJ Jr, Gorelick PL, Nagashima K, Gonda MA, Gilden RV, Tully JG, Russell RJ, Benveniste RE, Paster BJ, Dewhirst FE, Donovan JC, Anderson LM, Rice JM. 1994b. Chronic active hepatitis and associated liver tumors in mice caused by a persistent bacterial infection with a novel Helicobacter species. J Natl Cancer Inst 86:1222-1227.
Weisbroth SH. 1995. Pneumocystis carinii: Review of diagnostic issues in laboratory rodents. Lab Anim 24:36-40.
Weisbroth SH, Paganelli RG, Salvia M. 1977. Evaluation of a disposable water system during shipment of laboratory rats and mice. Lab Anim Sci 27:186-194.
Wescott RB. 1982. Helminths. In: Foster HL, Small JD, Fox JG, editors. The Mouse in Biomedical Research. Vol II. Diseases. New York: Academic Press. p 373-384.
Yang FC, Paturzo FX, Jacoby RO. 1995. Environmental stability and transmission of rat virus. Lab Anim Sci 45:140-144.
Zweig MH, Robertson EA. 1987. Clinical validation of immunoassays: A well-designed approach to a clinical study. In: Chan DW, Perlstein MT, editors. Immunoassay, A Practical Guide. Boston: Academic Press. p 97-127.
TABLE 1 Examples of sample selection errors
| Result | Methodology | Error |
| False-negative | Serology | Acutely ill, serum antibodies not yet detectable Immunodeficient or immunosuppressed, weak or no antibody response |
| Bacteriology/parasitology | Older and recovered from infection Site where organism is not normally resident | |
| All | Small sample size Sentinels not adequately exposed via soiled bedding or contact to infectious agents carried by principals | |
| False-positive | Serology | Strain with autoimmune disease Immunized or inoculated with biological material (such as tumor cells) Maternal antibodies |
| All | Sentinels housed under less strict conditions than principals (such as principals kept in microisolation cages, but sentinels are in open cages) |
| 1. Appropriately anesthetize the animal to collect blood for a serum sample. Anesthetize each separate sample group in an initially sterile anesthetic chamber so that respective results apply strictly to that particular group. If needed, obtain a blood film to check for the hemoprotozoa Eperythrozoon and Hemobartonella. 2. Record the body weights for each animal. 3. Examine the carcass grossly for lesions suggestive of dermatophytosis, ectoparasitism, and other causes of dermatosis and alopecia. Check the pelage and skin directly by low power dissection microscopy for mite and louse genera that include Polyplax, Myobia, Myocoptes, and Radfordia. Detection of mite genera that include Sarcoptes, Notoedres, Demodex, and Psorergates also requires microscopic examination of skin scrapings. 4. Sample the nasoturbinates by a wash or swab for retrieval of bacterial and mycoplasmal agents associated with the upper respiratory tract, including Streptococcus pneumoniae, Pasteurella pneumotropica, Staphylococcus aureus, C. kutscheri, Bordetella bronchiseptica, and Mycoplasma pulmonis. 5. Obtain an oropharyngeal sample (if needed) by swab for potential isolation of bacterial forms favoring that location (P. aeruginosa, Klebsiella pneumoniae and K. oxytoca, beta hemolytic Streptococcus, and Streptobacillus moniliformis. 6. Check the tympanic bullae (if routinely done, as in some laboratories) to aspirate any exudates for culture of bacteria colonizing the upper respiratory tract. 7. Reflect the skin from the ventral midline of the thorax and abdomen and, using scissors, remove the ventral body wall over those parts. Examine the thoracic organs grossly for lesions (abnormal appearance). Excise 1 or more pulmonary lobes for histological fixation or for extraction of DNA for pathogen genomic assay such as the polymerase chain reaction (PCR). Microscopic examination of lung sections is quite informative both for assessment of specific agents such as cilia-associated respiratory bacillus, C. kutscheri, M. pulmonis, and P. carinii as well as for anatomic correlation with infection by a number of other agents that may be detected by other modalities of the test battery (such as serology). These agents may or may not be on the profile list. 8. Examine the abdominal organs grossly for appearance and lesions that might prompt additional samples for histological fixation or microbial culture. In the absence of lesions, the critical tissues selected for histopathological examination include lung, ileum, liver, and kidney. The number of tissues chosen might be less or more, depending again on prosector preference, purpose of the screen, and economic considerations. Microscopic examination of sections of the ileum and liver are used to supplement serology, fecal culture, and PCR for a number of agents that include Salmonella, C. piliforme, and C. kutscheri. Histological sections of these organs may correlate with other tests in the battery or suggest other agents that are not on the profile list and prompt further examination. 9. Place wet mounts (slurries of intestinal scrapings) from the ileum under coverslips for potential detection of intestinal protozoa that include Spironucleus (Hexamita) muris, Giardia, Trichomonas, and Entamoeba. Some laboratories diagnose these agents alternatively by examination of stained sections of the intestines. 10. Resect the urocyst (if required) and examine for Trichosomoides crassicauda lying on the mucosa. 11. Resect the appendix---to lie with the contents and mucosa up---by opening with scissors and transecting along the long axis. Examine for pinworms that include Syphacia and Aspicularis under the dissecting microscope. 12. Collect fecal samples for (a) aerobic culture of enteric pathogens that include Salmonella, Citrobacter rodentium, and Helicobacter; (b) DNA extraction for a number of enteric pathogens that include Helicobacter sp., C. piliforme, and C. kutscheri; and (c) fecal flotation to check for coccidia (Eimeria) and Hymenolepis ova, as well as to act as a backup check for pinworms and a number of more rare helminths. Distribute the serum samples, inoculated microbiological media, tissues fixed for histological preparation, tissue and/or fecal samples for genomic analysis, and fecal samples to the appropriate laboratories where the in vitro process of diagnostic detection and identification of microorganisms begins. |
| Viruses | Parasites | ||||
| Microbiological Category | Husbandry | Competentc | Deficientc | ||
| Gnotobiotic | Isolator | . | . | . | . |
| . | Microisolation caged | . | . | . | . |
| SPF | Isolator | . | . | . | . |
| . | Microisolation cagec | . | . | . | . |
| . | Open cage-barrier | . | . | . | . |
| . | Open cage-nonbarrier | . | . | . | . |
a
Rare | . | Infrequent | . | Occasional | . | Common | . | Not Applicable | . |
c Immune status of host.
d Used according to recommended practices, that is, disinfected supplies and changed in a laminar floor hood.
FIGURE 2. Effects of microbiological category, husbandry, and host immune status on the chances of significant contamination with viruses, bacteria, and parasites.
(No.) | |||||||
| aRB (4) | |||||||
| 8-12 wk (4) | |||||||
| 4-5 wk (4) | |||||||
bCorynebacterium kutscheri.
FIGURE 3 Sample rat protocol: comprehensive health monitoring.
| Step no. | Protocol | Offset (no. of wk)a |
| Comprehensive testingb | ||
| Serology only | ||
| Serology only | ||
| Comprehensive testing |
bIncluding serology, bacteriology, pathology, and parasitology.
FIGURE 4 Sample schedule template.
Colony: | X | ||
Species: | Rat | ||
Start Date: | 1-Jan-98 | ||
| Comprehensive testing | 1-Jan-98 | ||
| Serology only | 29-Jan-98 | ||
| Serology only | 26-Feb-98 | ||
| Comprehensive testing | 26-Mar-98 |
FIGURE 5 Sample colony schedule.
FIGURE 6 Comparison of typical and ideal serology test.
| Assay Result | Exposed | Unaffected |
| Positive | TP | FP |
| Negative | FN | TN |
FIGURE 7 Definition of assay sensitivity and specificity. TP/FP, true-/false-positive; FN/TN, false-/true-negative.
| Assay Result | Exposed | Unaffected |
| Positive | 9800 | 4500 |
| Negative | 200 | 85,500 |
| Assay Result | Exposed | Unaffected |
| Positive | 980 | 4950 |
| Negative | 200 | 94,050 |
FIGURE 8 Effect of prevalence on the predictive value of positive results (PV+) for an assay with a specificity of 80%.
Copyright © 2008. National Academy of Sciences.
All rights reserved.
500 Fifth St. N.W., Washington, D.C. 20001.
Terms of Use and Privacy Statement