PR8 Influenza Virus Research Articles

Virus-neutralizing antibodies of immunoglobulin G (IgG) but not of IgM or IgA isotypes can cure influenza virus pneumonia in SCID mice

The ability of monoclonal antibodies (MAbs) to passively cure an influenza virus pneumonia in the absence of endogenous T- and B-cell responses was investigated by treating C.B-17 mice, homozygous for the severe combined immunodeficiency (SCID) mutation, with individual monoclonal antiviral antibodies 1 day after pulmonary infection with influenza virus PR8 [A/PR/8/34 (H1N1)]. Less than 10% of untreated SCID mice survived the infection. By contrast, 100% of infected SCID mice that had been treated with a single intraperitoneal inoculation of at least 175 micrograms of a pool of virus-neutralizing (VN+) antihemagglutinin (anti-HA) MAbs survived, even if antibody treatment was delayed up to 7 days after infection. The use of individual MAbs showed that recovery could be achieved by VN+ anti-HA MAbs of the immunoglobulin G1 (IgG1), IgG2a, IgG2b, and IgG3 isotypes but not by VN+ anti-HA MAbs of the IgA and IgM isotypes, even if the latter were used in a chronic treatment protocol to compensate for their shorter half-lives in vivo. Both IgA and IgM, although ineffective therapeutically, protected against infection when given prophylactically, i.e., before exposure to virus. An Fc gamma-specific effector mechanism was not an absolute requirement for antibody-mediated recovery, as F(ab')2 preparations of IgGs could cure the disease, although with lesser efficacy, than intact IgG. An anti-M2 MAb of the IgG1 isotype, which was VN- but bound well to infected cells and inhibited virus growth in vitro, failed to cure. These observations are consistent with the idea that MAbs of the IgG isotype cure the disease by neutralizing all progeny virus until all productively infected host cells have died. VN+ MAbs of the IgA and IgM isotypes may be ineffective therapeutically because they do not have sufficient access to all tissue sites in which virus is produced during influenza virus pneumonia.

Presentation by a major histocompatibility complex class I molecule of nucleoprotein peptide expressed in two different genes of an influenza virus transfectant.

Major histocompatibility (MHC) class I glycoproteins are specialized to present to CD8+ T cells, peptides that originate from proteins synthesized within the cytoplasm. Conventional killed vaccines are unable to get into the cell cytoplasm and therefore fail to expand the CD8+ T cell population. We have created a novel influenza transfectant virus, R10, which carries an immunogenic peptide from the nucleoprotein (NP) of PR8 influenza virus in its hemagglutinin (HA) and another similar peptide in its HK influenza virus NP. The two peptides are both presented by H-2Db and bind with approximately equal affinity. They can compete with one another for binding to H-2Db. Yet in cells infected with R10, both peptides are presented efficiently enough to expand the respective cytotoxic T lymphocyte (CTL) precursors in vivo and to serve as targets for CTL lysis in vitro. It has been proposed that proteins bearing signal sequences may be processed by a transporter-independent pathway. To investigate this, we infected the transporter-deficient cell line RMA-S with the R10 virus to see if the NP peptide expressed by the HA would be presented. The result shows that even the presence of a signal peptide in the HA does not overcome the lack of a transporter function, suggesting that the presentation of both peptides is dependent on functional transporter proteins. Our data also suggest the feasibility of creating by genetic engineering, recombinant vaccines expressing multiple epitopes that can effectively stimulate a cellular immune response.

Low avidity recognition of a class II-restricted neo-self peptide by virus-specific T cells.

The specificity with which CD4+ T cells recognize self peptides in vivo was examined in transgenic mice that express an influenza virus PR8 hemagglutinin (HA) polypeptide in many tissues, including the thymus (HA Tg mice). HA Tg and non-Tg mice were analyzed for their T cell responses to the major PR8 HA I-E(d)-restricted CD4+ T cell determinant S1. Negative selection eliminated S1-specific T cells from HA Tg mice. Nevertheless, HA Tg mice retained the ability to mount a T cell response to a closely related analog of the S1 determinant [S1(K113)], and some S1(K113)-specific TCRs displayed a partial reactivity with S1 as indicated by their ability to transmit signals for IL-3 but not IL-2 secretion in response to the neo-self peptide. Moreover, the neo-self S1 peptide antagonized the ability of these TCRs to signal IL-2 secretion in response to the foreign S1(K113) determinant. Thus, TCRs that exhibit a partial reactivity with a self peptide are present in the peripheral T cell repertoire and can be activated by a virus containing an analog of the self peptide. These findings provide a model for the induction of autoimmunity by viruses that are close homologs of self peptides, and suggest a way in which TCRs could react with self peptides during positive selection of developing thymocytes.

Liposomes and influenza viruses as an in vitro model for membrane interactions II. Influence of vesicle size and preparation methods

As previously demonstrated, membrane fusion and transfer of prodrug-type molecules between membranes can be studied with the octadecylrhodamine B chloride (R18) marker. We use this assay with PR8 influenza viruses and small (SUV) or large (LUV) unilamellar liposomes containing the virus receptor G D1a as interacting partners in order to analyse the effect of membrane curvature on fusion and transfer. Computer analysis revelaed no difference in the kinetics of fusion between PR8 with large or small liposomes respectively. Lipid transfer, however, is about 5 to 6 times faster from SUV than from LUV. We also tested the influence of the production method for liposomes, namely detergent dialysis, sonication and extrusion on membrane interactions. Fusion is not affected by either, whereas transfer is slower with liposomes produced by sonication.

Liposomes and influenza viruses as an in vitro model for membrane interactions I. Kinetics of membrane fusion and lipid transfer

To study membrane interactions, we use a previously described model system with intact PR8 influenza viruses and small liposomes containing the virus receptor G D1a. The assay is based on the fluorescent membrane marker octadecylrhodamine B chloride, R18, incorporated in liposomes. With this assay two processes can be studied in parallel: hemagglutinin-dependent virus/liposome fusion and spontaneous marker transfer. We present a detailed kinetic analysis of fusion and transfer which takes into account the contribution of dequenching in donor and acceptor vesicles, respectively. Both fusion and transfer follow second-order kinetics; they differ, however, in the interacting species. Fusion is collision-mediated, i.e. it depends on the total particle concentration. For R18 transfer, the rate is independent of the particle concentration, but determined by the R18 surface density, i.e. the R18/lipid ratio, which is responsible for the initial quenching of the labelled species.

Inactivation of PR8 influenza virus through the octadecylrhodamine B chloride membrane marker.

The octadecylrhodamine B chloride (R18) membrane marker was incorporated into PR8 influenza viruses and virus receptor (GD1a-) containing small unilamellar vesicles (SUV). Both were tested in a fusion/lipid transfer assay [Wunderli-Allenspach, H., & Ott, S. (1990) Biochemistry 29, 1990-1997] to find out whether incorporation into artificial and biological membranes yields equivalent results. The R18 assay is based on incorporation of quenched concentrations of the label into donor membranes and monitoring of the dequenching upon its dilution into unlabeled acceptors. With PR8 viruses and R18-labeled SUV, a fast, hemagglutinin-specific fusion takes place at pH 5.3, independently of the initial quenching. At neutral pH, a slow, nonspecific R18 transfer occurs. Both processes follow second-order kinetics. Upon incubation of R18-labeled PR8 viruses with unlabeled SUV or LUV in neutral buffer, transfer is also found. At pH 5.3, a complex dequenching curve best described with superposition of two second-order functions was encountered: a fast, hemagglutinin-specific component and a slow, nonspecific component. A decreasing proportion of the fast fusion was found with increasing initial quenching of labeled virus. Extrapolation showed that full fusion activity is obtained only with low initial quenching (20-30%). The gradual inactivation of the virus by increasing amounts of R18 was confirmed with biological assays (e.g., infectivity). The R18 surface density is much higher in viruses than in liposomes to obtain the same initial quenching. Analysis with the Stern-Volmer plot revealed that R18 monomers and dimers contribute to the quenching in labeled PR8 viruses, whereas only dimers determine the quench curve in liposomes.

Induction of Antihemagglutinin Antibodies by Polyclonal Antiidiotype Antibodies

Antiidiotypic antibodies can be envisioned as an alternative approach in the development of vaccines against influenza virus, which exhibits natural antigenic variations. In our work, we obtained two polyclonal cross-reactive anti-Id antibodies against PY102, VM113, and VM202 mAbs, which in turn are specific respectively for PR8 virus and laboratory-induced virus variants (PY102-V1 and VM113-V1). With these cross-reactive anti-Id antibodies, we were able to elicit anti-HA antibodies in mice. In comparing the anti-HA antibody response in animals injected with anti-Id antibodies to those immunized with PR8 influenza virus, we demonstrated that the HI titer was higher after virus immunization and that the PR8 virus boost was more efficient in this group. Our results showed that the polyclonal cross-reactive anti-Id antibodies were more efficient than the individual anti-Ids at eliciting responses. At the same time, we demonstrated that PR8-primed T cells, cultured with B cells from animals immunized with anti-Id antibodies, were able to produce anti-PR8 antibodies subsequent to stimulation with influenza virus.

A sensitive method to detect defined peptide among those eluted from murine MHC class II molecules

We developed a sensitive competitive inhibition radioimmunoassay able to trace pmoles of a defined peptide eluted from major histocompatibility complex (MHC) class II molecules that were subsequently fractionated by RP-HPLC. In this assay we used a model synthetic peptide corresponding to amino acid residues 110–120 from the hemagglutinin (HA) of PR8 influenza virus, and affinity purified rabbit antibodies specific for this peptide. The HA110–120 peptide binds to I-E d class II molecules on the surface of APCs and is recognized by specific CD4 + T helper cells. 2PK3 B lymphoma cells (H-2 d) were pulsed with HA110–120 peptide or PR8 virus, lysed, the MHC class II molecules extracted, and bound peptides eluted. After separation by RP-HPLC, the fractions were tested for inhibition of the binding of rabbit anti-HA110–120 antibodies to peptide coated microtiter plates. A significant inhibitory activity was observed with one peak when the cells were pulsed with HA110–120 peptide and two peaks when pulsed with PR8 virus. The inhibitory activity was correlated with the presence of HA110–120 peptide as demonstrated by peptide sequencing. The assay is reproducible and sensitive to 1 pmol of antigenic peptide. This assay can be useful to identify microbial peptides with defined structure and antigenicity among the multiple peptides bound to class II molecules.

Cytotoxin production by a merineLyngbya strain (cyanobacterium) in a large-scale laboratory bioreactor

A cytotoxic compound was produced by the marine cyanobacteriumLyngbya sp. Pearl strain in large laboratory-scale batch cultures. Adsorption and fractionation of methanol extracts with reverse phase (C-18) cartridges provided a rapid method for removal of bioassay interference from salts, biopolymers and pigments and concentration of the cytotoxic principles. Cytotoxicity to the murine leukemia cell line P-388 was produced in two cycles coinciding with the initiation of exponential growth and again during the late exponential growth phase. Antiviral activity against influenza virus PR8 was found in extracts prepared from early exponential growth phase cells but antiviral activity was not detected in extracts of mid-log or late-log growth phase cells. These differences in bioactivity suggests that the cytotoxic principles produced during early and late exponential growth may be different compounds. Cytotoxicity assays using murine P-388 leukemia indicates that the semi-pure compound has an IC50 of < 0.25 μg ml−1 to this cell line. P-388 cytotoxicity in cell extracts increased during the late exponential growth phase and the specific yield was estimated at approximately 0.14 mg g−1 (dry cells).

Passive transfer of local immunity to influenza virus infection by IgA antibody.

Secretory IgA is presumed to be the mediator of mucosal immunity based on many studies that show a correlation between protection and secretory IgA titers; however, a causal relationship has not yet been established. Classically, passive transfer of antibody has been used to demonstrate causality, but the passive transfer of local immunity with physiologically transported IgA has not been previously reported. In this study mice were injected intravenously with polymeric IgA (pIgA), monomeric IgA (mIgA), or IgG1 mAb specific for the H1 hemaglutinin of PR8 influenza virus. pIgA was shown to be specifically transported into nasal secretions relative to the mIg. The transported pIgA was functional, as evidenced by its ability to bind to virus in an ELISA assay and to protect nonimmune mice against intranasal infection with H1N1 but not H3N2 influenza virus. Intravenous injection of similar virus-neutralizing doses of anti-influenza IgG1 mAb did not protect against nasal viral challenge. IgA-mediated protection could be abrogated by the intranasal administration of antiserum against the alpha chain of IgA. These data demonstrate the passive transfer of local immunity by the i.v. administration of pIgA antibody and show that the IgA in secretions can protect against influenza virus infection. This general approach could provide a model for the evaluation of the role of local IgA in host defense against other pathogens.

Detection of toxic viral-associated double-stranded RNA (dsRNA) in influenza-infected lung

While many of the molecular events in viral replication are well studied, the molecular mechanisms by which viral infections trigger such constitutional symptoms as fever and ‘malaise’ are unknown. The hypothesis that these viral constitutional symptoms can be triggered by the toxic action of dsRNA associated with viral replication was investigated. Total lung RNA from mice acutely infected with PR8 influenza virus, but not from sham-infected mice, was shown to induce fever and altered sleep (excess slow-wave sleep, enhanced amplitudes of electroencephalographic slow waves, and reduced rapid eye movement sleep) when injected into the rabbit brain. Viral-associated dsRNA was shown to be responsible for the rabbit responses by differential nuclease digestion. Influenza viral dsRNA was directly demonstrated in the active lung RNA preparations by reverse transcriptase-polymerase chain reaction techniques. The time course of the responses paralleled those seen in the same model inoculated with nanogram quantities of the synthetic dsRNA polyriboinosinic-polyribocytidylic acid and suggested that they were mediated by induced cytokines. A model for the role of viral-associated dsRNA in eliciting both local cytotoxicity and viral constitutional symptoms is presented.

Temperature-dependent kinetics of the activities of influenza virus

The temperature dependence of membrane interactions between PR8 influenza virus and virus receptor (G D1a)-containing liposomes was studied. For quantitation, the octadecylrhodamine B chloride (1118) membrane marker was incorporated into liposomes at quenched concentrations. Upon interaction with target membranes, the marker gets diluted, and dequenching can be measured in a fluorescence spectrophotometer. Rate constants were calculated from the dequenching curves under low pH conditions, which allow for fusion, and at neutral pH, where no specific fusion occurs. Activation energies were determined from Arrhenius plots. The results were compared with the temperature dependence of other viral activities like infectivity, hemolysis, and fusion with erythrocytes. For the slow reaction at pH 7.4, where only non-specific lipid transfer takes place, the activation energy was about 24 kcal/mole between 15°C and 45°C. For the fast, hemagglutinin (HA)-specific fusion reaction (pH 5.3), a very low activation energy (∼7 kcal/mole) was found between 25°C and 37°C, whereas below 25°C it was much higher (∼34 kcal/mole). The temperature range with low activation energy coincides with the one for optimal infectivity, hemolysis, and fusion with erythrocytes. Furthermore, it is the same range in which the conformational change of HA takes place, which in the absence of a partner membrane leads to an irreversible inactivation of the fusion protein.

Kinetics of fusion and lipid transfer between virus receptor-containing liposomes and influenza viruses as measured with the octadecylrhodamine B chloride assay

Octadecylrhodamine B chloride (R18) and ganglioside GD1a (virus receptor) were incorporated into small unilamellar liposomes [Hoekstra et al. (1984) Biochemistry 23, 5675-5681]. Upon interaction of these liposomes with PR8 influenza viruses without prebinding, two types of dequenching were observed at 37 degrees C, both second-order processes: a fast reaction at pH 5.3, 2k = 17.53 x 10(-3) (Q.s)-1, and a slow reaction at pH 7.4, 2k = 0.335 x 10(-3) (Q.s)-1. The maximal level of dequenching was the same for both. Upon prebinding of liposomes to PR8 viruses (30 min, 0 degrees C, pH 7.4) at high concentrations, a very fast dequenching occurred when the prebinding mixture was diluted into prewarmed (37 degrees C) 10 mM PBS, pH 5.3. For the initial phase, a first-order rate constant of 0.5 s-1 could be extrapolated. After a quick drop in velocity during the first 30 s, the reaction was kinetically indistinguishable from the one found without prebinding. A second-order process with 2k = 16.52 x 10(-3) (Q.s)-1 became rate-limiting. The fast reactions at pH 5.3 can be abolished by inactivation or removal of the virus hemagglutinin. We conclude that the reaction at pH 5.3 reflects the hemagglutinin-dependent fusion process known to occur between influenza viruses and partner membranes at low pH; however, second-order kinetics indicate that specific binding rather than fusion is the rate-limiting step. For the slow dequenching, which is not affected by prebinding, the rate constant is 20 times lower than for the fast reaction, and the process is independent of viral hemagglutinin.(ABSTRACT TRUNCATED AT 250 WORDS)

Nucleotide changes in sequential variants of influenza virus hemagglutinin genes and molecular structures of corresponding monoclonal antibodies specific for each variant.

We have generated four monoclonal antibodies specific for one or more members of a series of two sequentially derived PR8 influenza virus variants. Three of these antibodies share cross-reactive idiotypes. The amino acid sequences of these antibodies were determined, and it was found that two of these antibodies use genes from the VH7183 family, whereas the third uses a gene from the VHJ558 family. All four monoclonal antibodies derive from different families of genes encoding the variable region of the kappa chain. The RNA sequence of the parent PR8 virus as well as the RNA sequences of the sequential variants were also determined, and it was demonstrated that the variant hemagglutinin molecules differed from the parent molecule by only a single amino acid interchange. Despite these subtle differences in antigenic structures of hemagglutinin, and the cross-reactive idiotype of the antibodies, their primary structures were very different. These data reinforce the idea that a wide variety of antibody structures exist which are directed against subtly different structures in biologically important antigens.

Virus infection as a stressor: Influenza virus elevates plasma concentrations of corticosterone, and brain concentrations of MHPG and tryptophan

Balb/c mice were infected with influenza virus PR8 (H1N1) by the intranasal route. At various subsequent times, brain samples were examined for their content of catecholamine and indoleamine metabolites, and plasma corticosterone was measured. Virus infection was associated with a progressive loss of body and thymus weights, and an increase in plasma corticosterone. Spleen weight initially increased then decreased. There were also increases in the cerebral content of free tryptophan throughout the brain, and of MHPG, a major catabolite of norepinephrine, especially prominent in the hypothalamus. Thus influenza virus can be regarded as a stressor because, like behavioral stressors, it activates the hypothalamic-pituitary-adrenal axis, and increases cerebral concentrations of tryptophan and norepinephrine catabolites. These changes resemble those observed following administration of sheep red blood cells and Newcastle disease virus, noninfectious activators of the immune system, suggesting that noradrenergic and HPA activation are common concomitants of antigenic stimulation. The mediator of these effects may be interleukin-1 released by activated macrophages. It should be noted that animals infected with viruses can be expected to exhibit stress-like endocrine and neurochemical changes.

Methods for the selection and growth of antigen-specific cytolytic T lines and clones bearing a defined T cell receptor β chain marker

Murine cytolytic T lymphocyte (CTL) clones specific for type A influenza virus antigens were generated by in vitro stimulation with syngeneic virus-infected cells in the presence of T cell growth factor (TCGF). All CTL clones recognize viral determinants shared by PR8 and X31 influenza viruses in association with a class I antigen, coded either by the H-2K or H-2D end of the appropriate haplotype. All clones express the Lyt2 antigen marker. Two of five clones also express an antigenic determinant of the Vβ chain of the T cell receptor (TCR) identified by F23.1 monoclonal antibody. To effectively generate F23.1 + and antigen-specific CTL clones, heterogenous CTL lines were expanded with F23.1 coated Sepharose beads in the presence of TCGF and then stimulated with PR8 virus-infected cells. Thus, both the proliferative activity to PR8 and the expression of the F23.1 marker was increased significantly. Alternatively, F23.1 + T cells were sorted from in vivo primed mice and expanded with PR8 virus-infected stimulator cells in the presence of TCFG. This F23.1 + T cell line exhibited antigen-specific cytotoxicity for PR8 virus-infected target cells. Additionally, in an ‘FcR-focused killing’ assay only the F23.1 + CTL line and F23.1 + clones lysed Fc receptor bearing target cells in the presence of F23.1 antibody. These findings indicate that antigen-specific and F23.1 + clones can be selected with high efficiency by alternating stimulation with influenza virus-infected cells and with F23.1-coated Sepharose beads or through the use of a cytofluorograph. The usefulness of antigen-specific and F23.1 + CTL clones and other posbible strategies for their selection are discussed.

Effect of syngeneic anti-idiotypic antibody on influenza virus neuraminidase antibody response.

Influenza viruses possess two major surface glycoproteins - hemagglutinin (HA) and neuraminidase (NA). Py203, a monoclonal antibody (Ab) specific for the neuraminidase of the PR8 (H1N1) influenza virus, was used to prepare syngeneic monoclonal anti-idiotypic (anti-Id) Abs. From a BALB/c mouse immunized with Py203 (anti-N1), we obtained RM1, a monoclonal anti-Id Ab. The Py203-Id was detected in a significant fraction of immunoglobulins (Igs) in the primary and secondary responses elicited by PR8 (H1N1) and X31 (H3N2) viruses. In animals injected with minute amounts of RM1 and subsequently boosted with an identical dose of RM1, no detectable anti-NA activity was noted, but a significant increase in Py203-Id-bearing Igs was observed. In the sera of animals injected with minute amounts of RM1 and subsequently boosted with PR8 (H1N1) or X31 (H3N2) viruses, an increase in anti-NA activity and in the level of Py203-Id was noted. Animals injected with large amounts of RM1 and boosted with PR8 and X31 showed a marked suppression of the Py203-Id but no alteration in the anti-NA response. The anti-Id recognizes an idiotope (the Py203 idiotope) shared by antibodies specific for the N1 and N2 neuraminidase variants.

Shared idiotopes among antibodies encoded by heavy-chain variable region (VH) gene members of the J558 VH family as basis for cross-reactive regulation of clones with different antigen specificity.

A wide idiotype cross-reactivity was observed among six groups of monoclonal antibodies specific for arsonate and nitrophenyl haptens, hemagglutinin of PR8 and X31 influenza viruses, dextran, A48-idiotype, and a set of six monoclonal antibodies with unknown antigenic specificity. All of these antibodies are encoded by heavy-chain variable region (VH) genes belonging to the J558 VH family. This idiotypic cross-reactivity was determined by studying the binding of these antibodies to a panel of six monoclonal anti-idiotype antibodies, each one raised against a member of the six groups of monoclonal antibodies. The administration at birth of two such monoclonal anti-idiotype antibodies induced a long-lasting suppression not only of the corresponding idiotype but also of VH-related idiotypes with different antigenic specificities. These results suggest that the idiotypes encoded by VH genes that belong to the same VH gene family are interactive one with another. The possible physiological consequences of this immunochemical cross-reactivity are discussed.

Influenza-specific suppression: contribution of major viral proteins to the generation and function of T suppressor cells.

Suppressor cells were generated in BALB/c mice by two sequential injections of PR8 influenza virus (A/Puerto Rico/8/34[H1N1]) and were tested for their ability to inhibit proliferative cellular responses towards multiple viral and nonviral antigens. In this way, suppression specific to PR8 as compared with purified protein derivative (PPD) and keyhole limpet hemocyanin (KLH) antigenic responses was illustrated. Experiments involving adoptive transfer of suppression to naive hosts with subfractionated lymphocyte populations demonstrated that the suppressors were Lyt-2+ T cells. Two major questions were addressed with this system. First, a determination was made of which anti-viral protein proliferative responses were affected by the PR8-induced T suppressor (Ts) cells. Ts cells were found to inhibit proliferating cells with specificities for isolated hemagglutinin (HA), neuraminidase (NA), nucleoprotein (NP), and matrix (M) antigens. Second, experiments were conducted to analyze the viral proteins contributing to the induction of PR8-specific Ts cells. Inoculations with either isolated HA or a combination of M + NP proteins induced T suppression specific to proliferative responses towards PR8. These experiments illustrate the contribution of external (HA and NA) as well as internal (M + NP) viral proteins to Ts cell generation and function.

B cell repertoire diversity to PR8 influenza virus does not decrease with age.

The affect of aging on the B cell repertoire was investigated at the clonal level by studying the response of 24- to 26-mo-old BALB/c mice to the PR8 influenza virus (H1N1). By using the splenic fragment culture technique, it was found that the frequency of B cells specific for either the intact virion or the HA does not change with age. The anti-HA monoclonal antibodies obtained from culture were additionally analyzed for fine specificity by binding in RIA to a panel of six H1 variant viruses. This analysis showed a considerable similarity in the distribution and diversity of reactivity patterns between monoclonal antibodies derived from splenic B cells of young vs aged mice. These findings indicate that repertoire expression per se may not be truncated in aged mice, and imply that reductions in the diversity of serum antibodies in aged mice may be due to environmental regulatory mechanisms.