BALB/cJ-MIETC. If this virus is responsible for the leukemia in HRS/J mice,

Transcription

1 GENETIC CONTROL BY THE hr-locus OF SUSCEPTIBILITY AND RESISTANCE TO LEUKEMIA* BY H. MIErER, D. D. MYERS, AND R. J. HUEBNER THE JACKSON LABORATORY, BAR HARBOR, MAINE, AND VIRAL CARCINOGENESIS BRANCHY NATIONAL CANCER INSTITUTE, NATIONAL INSTITUTES OF HEALTH, BETHESDA, MARYLAND Communicated April 22, 1969 Abstract.-Leukemia was observed in 45 per cent of hairless mice at 8 to 10 months of age and 72 per cent at 18 months. The incidence of leukemia in normal heterozygous mice of the same age is about one per cent at 10 months and increases to 20 per cent at 18 months. Other tumors are rare; two mammary adenocarcinomas occurred in heterozygotes, and an epidermal squamous cell carcinoma was found in a hairless mouse. fmurine leukemia virus was isolated from normal and leukemic mice of both genotypes on SWR/J- and C57L/J-M\ETC, but not BALB/cJ-MIETC. If this virus is responsible for the leukemia in HRS/J mice, the mutant gene (hr) enhances susceptibility and the wild-type allele (+) induces resistance to leukemogenesis, i.e., malignant transformation of reticulo-endothelial tissues occurs rather than inhibition of viral replication as both genotypes harbor virus in high titers. The two types of HRS/J mice, h/hr and hr/+, are congenic, i.e., they differ only with respect to one allele (hrr or +) at the mutant locus. This single gene difference should lend itself to analysis of the interaction of a specific gene and murine leukemia virus. It is well established that genetic factors play a role in the etiology of spontaneous tumors in mice, but the mechanism of gene-controlled effects on susceptibility is unknown in both spontaneous or induced tumorigenesis. Several studies have shown that histocompatibility genes, notably H-2, are important in viral oncogenesis.1 5 Other work has implicated a locus, Tla, closely linked with H-2 in linkage group IX rather than H-2 itself.3 However, there is no doubt that the H-2 region influences the susceptibility to and affects the course of virus-induced oncogenesis.6 Among more than 300 mutant genes occupying more than 250 loci known in the mouse, certain non-h-2 loci have also been found to affect susceptibility to and biological activity of leukemia virus, specifically the Friend leukemia virus complex which initiates spleen focus formation. For example, alleles at the W- dominant spotting locus control the availability or quality, or both, of target cells for SFFV, and the mutant alleles at the Steel (Sl) locus confer refractoriness to SFFV Although specific associations of single genes with mammary and lung tumors have been reported, only one instance of an association of spontaneous leukemia and a specified gene, dilution (d), has been known thus far.9 In addition, the flexed-tail gene (f) which is responsible for transitory siderocytic anemia and vertebral fusions, directly influences susceptibility to methylcholanthrene-induced leukemia.10 We describe the genetic control of susceptibility and resistance to viral leukemogenesis by the hairless locus of linkage group III. 759

2 7607llMICROBIOLOGY: MEIER ET AL. PROC. N. A. S. Materials and Methods.-Animals: Hairless (gene symbol, hr) is an autosomal recessive mutation maintained in strain HRS/J inbred brother by sister for over 30 generations." All matings are normal female (hr/+) by hairless (hr/hr) male mice. The two types of mice are congenic, i.e., genetically they differ only with respect to one allele (hr or +) at the mutant locus or to very closely linked genes. Thus, any pathophysiological differences between them can be ascribed to the presence or absence of homozygosity for the hr-gene. The HRS/J strain is maintained in the MPS unit of the Jackson Laboratory. All steps in the multiplication and replacement of the mutants are carried out at this laboratory. About 70 pens of the HRS/J strain are "rotated out" or retired each month when the breeders are 8-10 months old. In the course of "rotation" procedures, Mr. Frederic Johnson, technical supervisor of the MPS, observed gross-"nodular" lesions in hairless but not haired mice. The "nodules" proved to be greatly enlarged subcutaneous lymph nodes, the lymphadenopathy being part of a leukemic process that generally involved lymph nodes and spleen to a greater extent than the thymus. The apparent difference in the occurrence and incidence of leukemia in hairless and haired mice at "rotation" age prompted a continuous survey for leukemia. Also, male and female retired breeders of both genotypes were held for aging in order to obtain data on longevity and rate of leukemia or tumor development with increasing age. For these purposes we selected animals which did not have gross evidence of leukemia at 8 to 10 months. The mice, separated by sex and genotype, were housed five per cage in stainless steel cages, 5 inches wide, 11 inches long, and 6 inches deep. Water and Old Guilford laboratory chow were provided ad libitum. Cages were changed at weekly intervals when fresh bedding and water were provided. Each mouse was individually identified and records were maintained on date of birth, date of death, and pathological findings. Once set out, the animals were not subsequently regrouped as their number decreased. Cages were checked daily for dead mice. Pathological findings: Autopsies were performed on all mice in holding and those from "rotation" in which gross diagnosis was uncertain. Tissues for microscopy were fixed in buffered neutral formalin, cut at 8,, and stained with hematoxylin eosin. Leukemic tissues from hairless mice and a mammary adenocarcinoma derived from a haired mouse were serially transplanted to both hairless and haired mice by subcutaneous trocar transplantations. Recipients were 4- to 6-week-old females. Also, in order to assure their genetic homogeneity with respect to major histocompatibility genes, reciprocal skin grafts were exchanged between 2-month-old hairless and haired mice of compatible sex. The mice were operated under sodium pentothal anesthesia and the grafts observed for 30 days. Complement-fixation (CF) tests for murine leukemia virus and virus isolations: Ten per cent extracts were prepared in Eagle's minimal essential medium with 250 units of penicillin and 250 meg streptomycin per ml, using Ten Broeck tissue grinders. The extracts were sonicated at setting 1 in a Heat Systems Co. sonicator with microtip attachment for 2 to 3 bursts, clarified by centrifuging at 2000 rpm for 20 min and used for CF tests and TC inoculations. CF tests for group specific antigen of murine leukemia virus were performed in both tissue extracts and inoculated TC (comul-test).'2 Primary mouse embryo TC were prepared as described by trypsinization of minced 15-day SWR/J, BALB/cJ, and C57L/J embryos.'3 Inoculations for virus isolations and assays for the comul-test were made into 1-day-old secondary plate cultures seeded with 350,000 cells in 60-mm dishes from Falcon Plastics, Inc. All cultures were maintained on Eagle's minimal essential medium in Earle's balanced salt solution containing 10% unheated calf serum with 2 mm glutamine, 100 units penicillin, and 100 mcg streptomycin per ml. They were incubated at 370C in a humidified cabinet with an atmosphere of 5% CO2 in air. Fluids were replaced twice weekly, and the cultures harvested at 21 days by scraping the cells into 0.6 ml of supernatant fluid per plate. This volume represents 1/20 of the supernatant fluid, and the suspension is referred to as 20 times (20X) cell pack. Half the cell harvest was frozen and thawed once and sonicated. Sonicates were tested

3 VOL. 63, 1969 MICROBIOLOGY: MEIER ET AL. 761 for MuLV/CF-antigen at initial dilutions of 1:2 and 1:4. All CF tests were performed by Mr. H. C. Turner of the Viral Carcinogenesis Branch, Serology Unit, National Cancer Institute. He employed the microtiter system, using 1.8 units of complement and 4 units of antisera as previously described.'4 Results.-Pathological characteristics of leukemia and solid tw ors: Leukemia was characterized grossly by lymphadenopathy, spleno- and hepatomegaly, and thymoma. The subcutaneous lymph nodes were palpable or visible in the unskinned animals. Their tumorous involvement was mostly generalized, although in a few instances the mesenteric and mediastinal lymph nodes were involved to the greatest extent. The lymph nodes usually were enlarged severalfold their normal size, cream-colored, homogeneous or with large follicles protruding as in the spleen. The spleen was up to five times its normal size, and the thymus was markedly enlarged as well (Fig. 1). %'>~~~~~~~~~~~~~~~~~~~~~~' elrebtteeis nethrlypadnopt no thmm. Th lyp node ar kaki FIG. 1.haracteristic features of lymphoid leukemia in a 1-month-old hr/hr are generalized lymphadenopathy, hepatosplenomegaly, and thymoma. FiG. 2.-Myeloid leukemia in hr/hr aged 17 months. Typically, the spleen is greatly enlarged, but there is neither lymphadenopathy nor thymoma. The lymph nodes are khakicolored. Microscopically diffuse lymphoid infiltration obliterated the normal architecture of spleen, thymus, and lymph nodes. The follicles of spleen andelymph nodes had either disappeared due to uniform lymphoid cell infiltration or were distinct and composed of slightly more mature neoplastic cells than the diffusely spreading lymphoid cells. Capsular, septal, and vascular invasions by leukemic cells were common. In fact, infiltration of peni- and paralymphatic organs occurred in most instances. The lymphoid infiltration was present in the peniportal areas of the liver and often so extensive as to strangulate the liver lobules. The sinuses were invaded to varying degrees with numerous small foci or lym-

4 762 MICROBIOLOGY: MEIER ET AL. PROC. N. A. S. phoid aggregations scattered throughout the liver tissue. Other organs, including kidneys and lung, were infiltrated as well. In the kidneys neoplastic tissue occurred in the cortex and cortico-medullary junction either in a diffuse or nodular pattern. In the lung, perivascular and peribronchial lymphoid cuffing was found. Whereas the predominant cell type of leukemia found in HRS/J mice up to retirement age was lymphocytic, animals one year and older showed a considerably greater variation at age of death from leukemia and in morphology of cell type. In fact, the majority of leukemias in both the homozygous mutant and heterozygous normal mice were myeloid in cell type. Characteristically, animals older than 15 months suffered from granulocytic leukemia, with high peripheral white blood cell counts of 100,000 per mm3 or greater and consisting predominantly of immature granulocytes. They showed mainly hepatosplenomegaly rather than thymoma and lymphadenopathy. The enlargement of the spleen was five to ten times its normal size. The lymph nodes were normal in size but khaki-colored (Fig. 2). Inwaddition to leukemia, we found two cases of mammary adenocarcinomas wh were of mainly alveolar type and one case of an epidermal tumor consisting of papilloma and squamous cell carcinoma. Tumor and tissue transplantation: Two series of trocar inocula were prepared from spontaneous leukemia of hairless mice, mainly from spleens and enlarged lymph nodes. No differences existed in transplantability of lymphoid leukemia between the mutant and heterozygous genotypes; no grafts were rejected. Reciprocal skin grafts exchanged between hairless and haired mice used to detect genetic differences other than at the hr locus were observed daily for 30 days. At the end of one month, the grafts looked healthy, and there was no evidence of rejection. Therefore, no major differences in histocompatibility genes existed as rejection usually occurs between 12 and 20 days for H-2 dissimilarities. Incidence of leukemia and solid tumors in HRS/J mice: At eight to ten months of ag,, 141 of 337 mutants had leukemia, whereas only four of 279 heterozygotes had 4iukemia. Therefore, the incidence of leukemia in homozygous mutants is 45 per cent and in heterozygotes slightly more than 1 per cent. All the leukemic mice of either genotype had typically lymphoid leukemia. With the exception of one mutant that died from lymphoid leukemia at age 13 months and five days, all animals dying from leukemia at older ages had myeloid leukemia. In these animals, hepatosplenomegaly was present without enlargement of visceral or peripheral lymph nodes and thymoma. Data on lifespan and leukemia prevalence are still being collected so that leukemia incidence cannot as yet be computed on a mean death age for the two genotypes. Seventy-two per cent of the mutants that had died up to 18 months of age had leukemia, whereas only 20 per cent of the heterozygotes had leukemia at death. In addition to leukemia, we found a cutaneous papilloma with areas of squamous-cell carcinoma in a hairless breeder, and two mammary adenocarcinomas occurred in breeding heterozygotes (Table 1). Complement-fixation tests and virus isolations: The incidence of MuLV/CF-

5 VOL. 63,1969 MICROBIOLOGY: MEIER ET AL. 763 TABLE 1. Incidence of leukemia in HRS/J. Leukemia Age incidence Solid tumors (months) Genotype (%) (numbers) 8 to 10* hr/hr 45 Cutaneous papilloma, squamous-cell carcinoma (1) hr/+ 1 Mammary adenocarcinoma (2) 18t hr/hr 72 hr/+ 20 * Some animals were retired at 8 months of age, and others at 10 months. The two months' delay in rotation was caused by an increased demand for mutants so that the breeding cycle was extended from 8 to 10 months. t The oldest living animals have reached 18 months of age. Data on lifespan and tumor incidence are still being collected. AG increased with age to 100 per cent in retired breeders of both genotypes. There were both positive and negative animals between one and seven months. The embryos were consistently negative. Embryos cannot be distinguished phenotypically, but as all matings of HRS/J are hr/+ female by hr/hr male, one half of the offspring are expected to be hr/+ and the other hr/hr (Table 2). The results of complement-fixation tests correlated well with those of the virus isolations. All CF-positive animals readily yielded virus; murine leukemia virus was isolated from both normal and leukemic mice of either genotype. Results are listed in Table 3. With one exception, the virus grew preferentially in SWR/Jand C57L/J-METC, but not in BALB/cJ-METC. Discussion.-We have observed lymphoid leukemia in 45 per cent of hairless mice at eight to ten months of age. About 72 per cent of these mice develop TABLE 2. CF results obtained with spleens from HRS/J of different ages. Normal/ Age leukemia Genotype CF results* 14 monthst Normal hr/hr + 14 months Normal hr/hr + 14 months Leukemia hr/hr + 14 months Leukemia hr/ months Leukemia hr/+ + 9 months Normal hr/+ + 7 months Normal hr/+ 5 months Normal hr/+ + 1 month Normal hr/hr 1 month Normal hr/hr 1 month Normal hr/hr + 1 month Normal hr/hr 1 month Normal hr/hr 1 month Normal hr/hr to 18-day embryos (5)T, hr/+ and hr/hr - 19-day embryos (5) hr/+ and hr/hr - 20-day embryos (5) hr/+ and hr/hr - * 3 + or greater reaction at 1:2 or higher dilutions. t Individual animals. I Mothers from which 3 litters of embryos were derived. Embryos cannot be distinguished genotypically, but since all matings of HRS/J are hr/ + female by hr/hr male, half the offspring are hr/ + and half hr/hr. (Number of fetuses per cesarian-derived litter.)

6 MICROBIOLOGY: MEIER ET AL. PROC. N. A. S. TABLE 3. Virus isolations from HRS/J mice. Mouse Embryo Tissue Culture Genotype inoculum 2 4 Plt P2 P3 PI P2 P3 P1 P2 P3 Tissue or CF results* SWR/J BALB/cJ C57L/J hr/hr Leukemia hr/hr Leukemia hr/hr Cutaneous papilloma or squamous-cell carcinoma Spleen hr/hr XB, spleen hr/hr XB, spleen XB, spleen hr/hr XB, spleen hr/+ hr/+ XB, spleen +... hr/+ XB, spleen hr/hr W, spleen 1:20 dilution hr/hr W, spleen 1:20 dilution hr/hr W, spleen : 10 dilution hr/+ W, spleen W, spleen... hr/+ + hr/+ W, spleen * 2 and 4, dilutions of MuLV-group-specific antiserum. t P1, TC passages. I XB, retired breeders (8 months or older). W, weanling (4 weeks of age). myeloid leukemia later on in life up to 18 months. The incidence of leukemia in normal, heterozygous mice is slightly more than one per cent at eight to ten months, and 21 per cent at a year and a half. Therefore, the incidence of leukemia in hairless is nearly 50 per cent greater than in haired mice, and the occurrence of leukemia is delayed in haired mice by about six months. We cannot as yet explain the interesting finding that lymphoid leukemia prevails in HRS/J mice up to 12 to 13 months of age and myeloid leukemia occurs in older animals. However, a similar situation occurs in man. Although chronic lymphoid and myeloid leukemia may appear at any age in man, the lymphatic type occurs predominantly in the first three decades (children and young adults), and the myeloid form tends to appear in the middle- or old-age groups. We have shown with the aid of tumor transplantation and reciprocal skin grafting that the two types of mice are congenic, i.e., genetically they differ only with respect to one allele (hr or +) at the mutant locus and are alike with respect to histocompatibility genes. Thus, the difference in leukemia occurrence and incidence can be ascribed to the presence or absence of homozygosity for the hrgene. We have isolated murine leukemia virus from both hairless and haired animals of the HRS/J strain on SWR/J- and C57L/J-METC, but not BALB/ cj-metc. If the virus is responsible for the leukemia in the hairless mice, the mutant gene enhances susceptibility and the wild-type allele induces delay and resistance to leukemogenesis. We are now producing homozygous normal animals (+/+) by mating heterozygotes and progeny testing. Homozygous wild-type animals may be entirely

7 VOL. 63, 1969 MICROBIOLOGY: MEIER ET AL. 765 resistant to leukemogenesis, and may or may not be virus- and CF-antigen positive. An assessment of the influence of allelic substitution on leukemia and tumor development and on the presence or absence of isolatable virus and CFantigen is expected to clarify the role of the alleles in viral oncogenesis. The influence of the hr-locus on leukemogenesis is of interest and importance for at least two other reasons: (1) the hr gene has previously been known to express itself only in the skin and to be functional only after the first hair cycle, and (2) ever since the discovery of the hr-mutation, the hairless mouse has been a favorite tool in cancer research. Specifically, the hairless mutant is useful for experiments involving application of substances directly to the skin. It is commonly used in cosmeto-pharmacological research and in studies of systemic and topical carcinogenesis.'5 The presence of murine leukemia virus and the high incidence of "spontaneous" leukemias in hairless mice may have influenced the results of such experiments and the evaluation of a test compound. The enhancing effect of chemical carcinogens on viral infection is well documented.'8 17 The effect of hr on leukemogenesis has not previously been described, although there has been a suggestion that hairless and other seemingly "irrelevant genes," such as vestigialtail, shaker-2, waved-2, fused- and flexed-tail genes, reduce body weight and decrease the incidence of lung tumors in inbred mice.18 We have found that the occurrence of the murine leukemia-sarcoma virus complex is widespread among inbred strains of mice, but that the incidence varies with strain, age, and whether or not the mice are tumor-bearing, confirming the earlier findings of Hartley et al.3 We cannot as yet explain these findings, but susceptibility to tumorigenesis is in part genetically determined as a threshold character, and certain genes can greatly modify virus-host cell interactions. The mechanism involved in these interactions have variously been termed "switch-on," activation or derepression of viral expression, and "switch-off" or repression.23 Our findings in tumored-hairless mice are comparable to "total switch-on," and the significantly decreased leukemia incidence in haired animals represents an instance of "partial switch-off." Although the concept of gene control and viral expression is a sound working hypothesis, nothing is known about the underlying mechanisms. We feel, however, that the single gene or allelic differences between the two types of HRS/J mice should lend itself to an analysis of the interaction of a specific gene and expression of murine leukemia virus. Since both genotypes harbor isolatable virus and CF-antigen in high titers, viral replication is probably similar, but malignant transformation of reticulo-endothelial tissue occurs more readily in the hairless than the haired genotype. Such an outcome of virus-cell or gene interactions has not previously been reported for RNA-tumor viruses of mammals. It is known that genetic factors are crucial in Rous sarcoma virus focus formation.' Also, human fibroblast cell strains are differentially susceptible to transformation by the oncogenic DNA containing SV40 and adenovirus Cell cultures of patients suffering from Fanconi's anemia are most markedly susceptible,2' and a number of genetic diseases of man are well known to be associated with an unusually high risk of developing tumors.22 The hrlocus in the HRS/J strain may be a model for and of considerable help in the elucidation of the genetic control and basis of susceptibility and resistance to

8 766 MICROBIOLOGY: MEIER ET AL. PROC. N. A. S. leukemia and other tumors. Determination of the factors influencing the outcome of murine leukemia virus cell interaction is of utmost importance for a better understanding of viral oncogenesis. Studies along these lines are planned. Abbreviations used: H-loci, histocompatibility loci; Tla-locus, thymus leukemia (antigen)- locus; SFFV, spleen focus-forming virus; MPS, mutant production stocks; CF, complement fixation test; TC, tissue culture; MuLV, murine leukemia virus; COMUL test, complement fixation test for MuLV; MuLV/CF-AG, murine leukemia virus complement-fixing antigen; METC, mouse embryo tissue culture. * This work was supported by U.S. Public Health Service contract PH from the National Cancer Institute. 1 Tennant, J. R., and G. D. Snell, Natl. Cancer Inst. Monograph, No. 22 (1966), p Lilly, F., Natl. Cancer Inst. Monograph, No. 22 (1966), p Lilly, F., Genetics, 53, 529 (1966). 4 Lilly, F., Proc. Am. Cancer Res., 8, 41 (1967). 5 Naredi, S., these PROCEEDINGS, 58, 485 (1967). 6 Tennant, J. R., and G. D. Snell, J. Natl. Cancer Inst., 41, 597 (1968). 7 Bennet, M., R. A. Steeves, G. Cudkowicz, and E. A. Mirand, Science, 162, 564 (1968). 8 Steeves, R. A., M. Bennet, E. A. Mirand, and G. Cudkowicz, Nature, 218, 372 (1968). 9 McDowell, E. C., J. S. Potter, and M. J. Taylor, Cancer Res., 5, 65 (1945). Law, L. W., J. Natl. Cancer Inst., 12, 1119 (1952). 1 Green, E. L., ed., in The Biology of the Laboratory Mouse (McGraw-Hill: New York, 1967), 2nd ed., 706 pp. 12 Huebner, R. J., these PROCEEDINGS, 58, 835 (1967). 13 Hartley, J. W., W. P. Rowe, W. I. Capps, and R. J. Huebner, these PROCEEDINGS, 53, 931 (1965); J. Virol., 3, 126 (1969). 14 Huebner, R. J., W. P. Rowe, H. C. Turner, and W. T. Lane, these PROCEEDINGS, 50, 379 (1963). 15 Homburger, F., A. Treger, J. R. Baker, and C. M. Crooker, Proc. Sci. Sect. Toilet Goods Assoc., 35, 5 (1961). 16 Lieberman, M., N. Haran-Ghera, and H. S. Kaplan, Nature, 203, 420 (1964). 17 Chieco-Bianchi, L., L. Fiori-Donati, L. G. DeBenedictis, and G. Tridenti, Nature, 199, 292 (1963). 18 Heston, W. E., J. Natl. Cancer Inst., 2, 127 (1941). 19 Payne, L. N., and P. M. Biggs, Virology, 24, 610 (1964). 20 Todaro, G. J., Natl. Cancer Inst. Monograph, 29, 271 (1968). 21 Todaro, G. J., and S. A. Aaronson, these PROCEEDINGS, 61, 1272 (1968). 22 Fraumeni, J. F., Jr., and R. W. Miller, J. Natl. Cancer Inst., 38, 593 (1967). 23 Huebner, R. J., J. W. Hartley, C. E. Whitmire, R. L. Peters, and H. Meier, in preparation.