Cell,
Vol. 13, 643-650,
April
1978,
Copyright
0 1978 by MIT
Two Plasma Membrane Antigens of Testicular Sertoli Cells and H-2-Restricted versus Unrestricted Lysis by Female T Cells Salvatrice Ciccarese and Susumu Ohno Division of Biology City of Hope National Medical Center Duarte, California 91010
Summary Sertoli cell-only seminiferous tubules of sterile XX, Sxr/-male mice served as an excellent source of pure Sertoli cells. When H-2-compatible female mice were immunized 3 times with these Sertoli cells, resulting antibodies recognized two antigens on the plasma membrane of testicular Sertoli cells. They were male-specific, but ubiquitously expressed H-Y antigen and the cell lineage-specific antigen which Sertoli cells shared with ovarian follicular cells. Doubly primed (2 or 3 times in vivo, and once in vitro) cytotoxic T cells from these females lysed target Sertoli cells in both H-2-restricted and nonrestricted manners. While H-2-restricted killings were attributable to H-Y antigen, further work is needed to identify the Sertoli follicular cell lineage-specific antigen as the cause of H-2-nonrestricted killings. Introduction A testis-organizing role has been assigned to the male-specific but ubiquitously expressed H-Y antigen (Wachtel et al., 1975b; Ohno, 1977; Wachtel, 1977). Inasmuch as testicular organogenesis involves not one but several cell types, however, it was felt that the possession of H-Y antigen by all cell types-while an absolute prerequisite-is by itself insufficient for testicular organogenesis. A set of additional plasma membrane antigens, each specific for a particular cell type, appeared essential since somatic elements within the central gonadal bastema of an emerging male embryonic gonad must sort themselves out before precursors of Sertoli ceils can get together to form the wall of testicular cords while not involving precursors of Leydig ceils in this process. Such sorting out processes are likely to depend upon the cell typespecific plasma membrane antigen through which those belonging to the same lineage can recognize each other. In vitro reaggregation experiments using mouse and rat gonadal cells have shown that testicular Sertoli cells lysostripped of H-Y antigen behaved as though they were follicular cells (Ohno, Nagai and Ciccarese, 1977; Zenzes et al., 1977). This indicated that testicular Sertoli cells and ovarian follicular (granulosa) cells belong to the same lineage; thus they might be marked by the identical plasma membrane antigen. One outstanding characteristic of H-Y antigen is
that cytotoxic female T cells treat it exactly in the same manner as a viral antigen, so that lysis by T ceils of H-Y-incompatible target cells is as H-2restricted in the mouse as lysis of virally infected or transformed target cells (Gordon, Simpson and Samelson, 1975; Gordon et al., 1976; Simpson and Gordon, 1977). As to the cause of H-2 restriction, direct evidence of physical association between HY antigen and &-microglobulin-HLA antigen dimers on the plasma membrane of male human cells has recently been obtained (Beutler et al., 1978), and equally direct evidence of physical association between a Friend leukemia viral antigen and /3,-microglobulin-H-2 antigen dimers on the infected mouse plasma membrane has been published (Blank and Lilly, 1977). One of us proposed that the original function of p,-microglobulin-MHC antigen dimers has been to serve as the plasma membrane anchorage sites of organogenesis-directing antigens such as testisorganizing H-Y antigen, and that for this reason, parasitic viruses evolved the strategy of displacing these functionally important host antigens with their own (Ohno, 197713). According to this proposal, plasma membrane antigens that are not organogenesis-directing are not in association with p,-microglobulin-MHC antigen dimers. We therefore asked, whether female T cell killings of testicular Sertoli cells are invariably H-2-restricted. Results Cytotoxic Specificities of Anti-Sertoli Antibodies Serum samples from H-2” as well as H-2” females immunized 3 times with H-Pcompatible XX, Sxrltesticular Sertoli cells were invariably a mixture of two antibodies: H-Y antibody and another antibody which detected the cell lineage-specific antigen shared by testicular Sertoli cells and ovarian follicular cells. The above was deduced by combinations of cytotoxicity and absorption tests as shown in Table 1. In the presence of agarose-absorbed guinea pig complements, unabsorbed antibodies were cytotoxic to male but not female epidermal cells (BALB/c), to ovarian follicular but not luteal cells (BALB/c), and to testicular Sertoli but not primordial germ cells (Sxr stock). A slight cytotoxicity to Leydig cells was apparently due to H-Y antibody. Although most cell types of the mammalian male readily absorb H-Y antibody raised in the mouse (Wachtel, Koo and Boyse, 1975a), this antibody is cytotoxic only to very limited cell types of male mice; epidermal cells are one such type (Scheid et al., 1972). When raw anti-Sertoli antibodies were a,bsorbed with an excess number of either male epidermal cells (7 x lo6 cells per 50 ~1 of ‘14 diluted antibod-
Cell 644
Table 1. Cytotoxic Specificities Sertoli-Follicular (S-F) Antigen
of Anti-Sertoli
Sera That Contained
H-Y Antibody
Testicular Primordial Germ
and Antibody
Directed
against
a Cell Lineage-Specific
Ovarian
Sertoli
Leydig
Epidermal
Follicular
Luteal
Male
Female
+++
-
-
+++
t
++
-
H-Y antibody removed by absorption with male BALB spleen or epidermal cells -
+++
-
++
-
-
-
-
-
-
Unabsorbed
Further cells
absorption
with
ovarian
(++) 21-27% corrected percentage corrected percentage killing.
follicular -
killing
(see
Figures
1 and
2); (++)
ies) or male spleen cells (12 x 10” cells per the same amount), male-specific cytotoxicity toward epidermal cells, as well as a slight cytotoxicity toward testicular Leydig cells, was totally abolished (Table l), while absorption with female epidermal cells had no effect on their cytotoxicity toward male epidermal cells (Figure 1). Once H-Y antibody was removed by absorption with the above noted male cells, what remained of the anti-Sertoli antibodies became highly specific; among testicular elements, they were cytotoxic only to Sertoli cells of all ages (5-100 postnatal days), while XY neonatal primordial germ cells and XX, Sxrl-. Leydig cells remained totally unaffected. Anti-Sertoli antibodies deprived of H-Y antibody were also cytotoxic to BALB/c ovarian follicular cells, but not at all to luteal cells (Table 1). Although antibodies were consistently more cytotoxic to testicular Sertoli cells than to ovarian follicular cells (Figure 2), we believe that the difference was attributable to a relative impurity of the ovarian follicular cell fraction. Compared with the uniform appearance of Sertoli cell preparations under the microscope, our ovarian follicular cell fractions contained considerable proportions of spindle-shaped cells and rectangular cells that were not killed by anti-Sertoli antibodies. The indication that the plasma membrane of testicular Sertoli cells and ovarian follicular cells possesses an identical or very similar antigenic determinant was further confirmed by the absorption test. When H-Y antibody-removed anti-Sertoli antibodies were absorbed with ovarian follicular cells at the level of 12 x lo6 cells per 50 PI of l/4 diluted antibodies, their cytotoxicity toward testicular Sertoli cells was completely absolished (Figure 2). In fact, doubly absorbed (first with male spleen or epidermal cells, and then with ovarian follicular cells) anti-Sertoli sera demonstrated no residual cytotoxicity to any of the cell types listed in Table 1. Thus the immunization of H-2-compatible female mice with XX,
15-20%
corrected
percentage
killing
(see Figure
l-5%
H-Zd Q Thrice Immunized With H-2d Sertoli Cells
1
ABSORSEO 9 EPIOERMQ? \ x
"NPSSORSEOX esSORsE0
2); (k)
-7
0
.
C~EPIOERMQL \
x .
\
0 x
i
----b CLmpieme”tP
.
b Cootro,
----_____--------_
4 ANTIFigure BALB ments
SERTOLI
6 CELL
1. Cytotoxicity of Raw Anti-Sertoli Epidermal Cells in the Presence
16 SERA
3’2 DILUTION
Antibodies of Guinea
toward Male Pig Comple-
This serum sample was absorbed with male and female BALB epidermal cells at the level of 2.5 x lo6 cells per 50 ~1 ‘/a diluted antibodies. Only male epidermal cells and male spleen cells removed the cytotoxicity. This revealed the presence of H-Y antibody within raw anti-Sertoli antibodies.
sxrl-testicular Sertoli cells apparently provoked the production of two antibodies: one directed against testis-organizing H-Y antigen, and the other directed against the cell lineage-specific antigen which marks testicular Sertoli cells and ovarian follicular cells as being derived from the same cell lineage. This is not to suggest, of course, that the above are the only antigenic determinants on the plasma membrane of testicular Sertoli cells. Indeed, xeno-
Membrane 545
Antigens
of Testicular
Sertoli
Ceils
H-Zd Q Thrice Immunized With
ANTI -SERTOLI
CELLS
Figure 2. Reaction of Anti-Sertoli Ovarian Follicular Cells
(-anti-HY) Antibody
SERA with
DILUTION
Sertoli
Cells and
Once H-Y antibody was removed by absorption, the remaining anti-Sertoli antibody was cytotoxic only to testicular Sertoli cells themselves and to ovarian follicular cells. The absorption by ovarian follicular cells at the level of 12 x lo6 cells per 50 PI ‘/4 diluted antibody completely removed its cytoxicity toward Sertoli cells. Thus it became clear that this antibody was detecting an identical or very similar antigenic determinant on the plasma membrane of testicular Sertoli cells and ovarian follicular cells.
genie antibodies such as rabbit anti-mouse Sertoli antibodies appear to recognize a great variety of antigens that are difficult to sort out. At any rate, the two presently identified appear most pertinent to testicular organogenesis. Needless to say, antiSertoli-follicular cell lineage-specific antibody was an autoantibody as far as immunized females were concerned. We obviously need to do an histological, as well as functional study of the ovaries of immunized females. Our impression, however, is that this cell lineage-specific antigen on the plasma membrane of follicular cells rapidly disappears after puberty, and it is for this reason that adult H2-compatible females were able to generate an antibody directed against this antigenic determinant. Whether the corresponding antigenic determinant found on XX, Sxr/- testicular sertoli cells can also be found on adult XY Sertoli cells has not been determined either. It is possible that in the absence of male germ cells, XX, Sxrl- Sertoli ceils can never attain the differentiated state characteristically postpubertal. Nevertheless, anti-Sertoli sera were cytotoxic to a crude fraction of juvenile XY Sertoli cells.
H-2 Restricted and Unrestricted Cells by Doubly Primed Female
Lysis of Sertoli T Cells
A highlight of this preliminary study on T cell killings of Sertoli cells was our direct observation of the actual killings under the microscope (Figure 5, right). Surprisingly, unprimed virgin female T cells of the control mixture readily attached themselves to the plasma membrane of Sertoli cells. Such attachments, however, caused no death of Sertoli cells, even when the large cell surface of a single Sertoli cell was completely covered by 40 or more small lymphocytes. In sharp contrast, doubly primed female T cells were extremely efficient killers in the H-2-restricted as well as unrestricted situation. Lysis of a Sertoli cell was often caused by the attachment of only one or two tiny T cells. Under the attack, a large Sertoli cell cytoplasm literally boiled as it became increasingly vacuolated. Then the nucleus suddenly became pale and swollen. An individual killing took no more than one hour. Even when there were 20 or more T cells in the close vicinity of one Sertoli cell, the actual killing appeared to have been done only by one or two T cells. The observed efficacy of cytotoxic T cells apparently was a primary reason for the fact that in spite of a great difference in their sizes, the maximal killing was already attained at the attacker-to-target ratio of 3:l (Figure 3). Female mice primed in vivo as well as in vitro with H-2-compatible male XY spleen cells generate cytotoxic T cells directed against H-Y antigen and lyse male target cells in the H-2-restricted manner (Gordon et al., 1975, 1976; Simpson and Gordon, 1977). One might expect such primed female T cells to recognize H-Y antigen on the plasma membrane of Sertoli cells as well, and to lyse them also in the H-2-restricted manner. Unfortunately, the above investigators have also shown that H-2” and H-2k females are least efficient in generating cellmediated immune responses against H-2-compatible male target cells. Most of the mated pairs of our Sxr stock were either H-2k or H-P homozygous; the H-2b subline just begun constituted a distinct minority. Thus we chose to immunize BALE/B6 (H2”lH-2”) F, females with male F, spleen cells in the hope that cytotoxic T cells that readily lyse not only H-2” male spleen cells, but also H-2” testicular Sertoli cells, could be generated. Gordon, Samelson and Simpson (1977), however, have shown that BALB/BlO (H-2”/H-2b) F, female T cells sensitized with male F, spleen cells readily lysed F, as well as those H-2b male spleen cells presented as their targets, but did not lyse H-2” male spleen cells. Accordingly, it is quite possible that had we presented H-2” male spleen cells as targets to our doubly primed H-2”/H-2b female cytotoxic T cells,
Cell 646
TARGET SERTOLl H-2d H-2k
ATT. T-CELLS H-2d/H-2b PRIMED IN VIVO-IN VITRO 0-O o-o UNPRIMED x-x
Table 2. H-2 Restricted Testicular Sertoli Cells
and Unrestricted by H-2 Compatibly Target
H-2d
l -------. \
Primed GSpleen
with:
Responder ? T Cells
Lysis of XX, Sxr/Primed Female T Cells Testicular
H-2 Type
Sertoli
Cells
H-2 Restriction
Lysis
Cells +++ - Yes
\
J
.
GSpleen
Cells ++ - Yes
H-2d/H-2b
“1 -__-x--------o------------~
Sertoli
Sertoli
\
Sertoli
ATTACKERS
3:l
: TARGETS
2:l
0:l
RATIOS
Figure 3. H-P-Restricted Lysis of XX, Sxr/-Testicular by H-2d/H-2b (BALB/BG) F, Female T Cells Doubly Male F, H-2d/H-2b Spleen Cells
Sertoli Primed
Cells with
H-2d but not H-2k targets were lysed. Target nuclei stained by trypan blue were scored as lysed cells. The attacker-to-target ratio of 3:l already attained the near-maximal killing of H-2-compatible targets. At the 4:l ratio, the experimental killing of H-Y targets was 47.5%, whereas the corresponding value by unprimed virgin H-p/H-2” female T cells was 23.8%. Thus the corrected percentage lysis of H-T Sertoli cells was calculated to be 31 .I%, whereas the corrected percentage lysis of H-2” Sertoli cells at the 4:l ratio was 0.8%. Thus in Table 2, The first value was scored as (+++), and the second value as (-). 250 ~1 of a reaction mixture always contained 1.5 x lo5 Sertoli cells as targets to which an appropriate number of doubly primed female T cells was added to the desired attacker-to-target ratio.
they might not have lysed them. Nevertheless, when H-2” testicular Sertoli cells were presented as the targets, they readily lysed them, while they were not at all cytotoxic to H-2k Sertoli cells (Figure 3 and Table 2). Thus it appeared that if directed against H-Y antigen, female T cell killing of testicular Sertoli cells is H-2-restricted. It should be pointed out that H-Y antigen sites are apparently far more prominently displayed on the plasma membrane of Sertoli cells than on that of male spleen cells. For the routine production of H-Y antibody, we give five weekly injections of 525 x IO6 male spleen cells each, yet only 25% of the immunized females yield H-Y antibody of usable strengths. In sharp contrast, only three consecutive injections of 1.5-2.0 x lo6 XX, SxrlSertoli cells each generated usable H-Y antibody in every instance as already noted. The difference, however, was confined to the frequency of success. When H-2” and H-2” females were doubly primed
H-2” H-T
H-2k
H-2” H-2d H-2k/H-2” H-2b
+++:
Yes
++t ++;
No +
Cells
H-2“
Sertoli
H-2* Cells
x------x
I
4:1
Cells
H-2*
H-2d H-2k H-2d/H-2k
++ ++ ++
No
H-F
H-P H-2b
:I
No
Cells
H-2” Attacker-to-target age lysis; (+++) 3); (++) 21-30% rected percentage
ratio (4:l). (++++) 41-46% corrected percent31-40% corrected percentage lysis (see Figure corrected percentage lysis; (+) il-20% corlysis; (3) 5-10% corrected percentage lysis.
(2 or 3 times in vivo, and once in vitro) not with male spleen cells but with testicular Sertoli cells, lyses by sensitized female T cells of target Sertoli cells were seldom H-2-restricted (Table 2). It is probable that on the plasma membrane of Sertoli cells, H-2-compatible female T cells recognized not only H-Y antigen, but also other antigenic determinants such as the Sertoli-follicular cell lineagespecific antigen which their B cells recognized, and that their cytotoxic effect upon target Sertoli cells, if directed against the latter antigen, was not H-2-restricted. Whether or not sensitizing Sertoli cells generate two or more kinds of cytotoxic female T cells can be tested by offering these T cells male spleen cells and ovarian follicular cells as targets. Lyses of these target cell types, however, have to be measured by the 51Cr release test. As already noted, we have thus far obtained unsatisfactorily low percentage lyses by this method. Nevertheless, if these Sertoli cell-sensitized female T cells lysed male but not female spleen cells in the H-2-restricted manner, there existed T cells directed against H-Y antigen. If these female T cells lysed ovarian follicular cells, this time in H-Bunrestricted fashion, cytotoxic T cells directed against the Sertoli-follicular cell lineage-specific antigen
Membrane 647
Antigens
were obviously killings of Sertoli
of Testicular
responsible cells.
Sertoli
for
Cells
H-Punrestricted
Discussion Although male-specific H-Y antigen is ubiquitously expressed by nearly all the cell types of mammalian males. At first glance, the assignment of a specific organogenesis role to an antigen so ubiquitously expressed might seem rather odd. Yet to direct organogenesis, the plasma membrane component responsible for it has to express itself before the commencement of that organogenesis. Indeed, in the development of mouse embryos, gonadal differentiation does not begin until the twelfth day (Peters, 1970). H-Y antigen, however, is first seen 10 days earlier in preimplantation male embryos of the 8 cell stage (Krco and Goldberg, 1976). The ubiquitous expression of H-Y antigen may be viewed as the natural consequence of a need to express itself so early in embryonic development. Nevertheless, for H-Y antigen (uniformly posessed by all the cell types of the male gonad) to direct testicular organization, it has to depend upon a set of cell type-specific antigens that enabie individual cells of the same lineage to identify each other. One such antigen found on the plasma membrane of Sertoli cells was also identified on ovarian follicular cells. Such sex-independent expression of the cell lineage-specific plasma membrane antigen by gonadal cells might be the very factor that enables H-Y antigen to play a pivotal role in the primary (gonadal) sex determination of mammals. Indeed, in vitro reaggregation experiments have shown that if lysostripped of their H-Y antigen, testicular Sertoli cells readily behaved as though they were ovarian follicular cells (Ohno et al., 1977; Zenzes et al., 1977). XX/XY gonads of experimentally produced blastocyst fusion chimaeras most often organize testes, less often ovaries and seldom ovotestes (McLaren, 1976). This suggests the presence of a development strategy that avoids the formation of mosaic organs. It has been suggested that this strategy is based on competitive displacements on the plasma membrane of one organogenesis-directing antigen with another, and that.such competitive displacements are made possible because competing antigens, such as testis-organizing H-Y antigen versus a yet to be identified ovary-organizing antigen, utilize the same plasma membrane anchorage site which is furnished by P,-microglobulin-MHC antigen dimers (Ohno, 1977b). Our previous finding on bovine freemartin gonads indicated a short-range hormone-like role of H-Y antigen (Ohno et al., 1976). Furthermore, XX cells that comprised 50% of the Sertoli, Leydig cell population in XX/XY chimaeric mouse gonads were found to be en-
dowed with as much H-Y antigen as XY cells (Ohno et al., 1978). The fact that female T cell killing of male target cells, if directed against H-Y antigen, is H-2-restricted may be viewed as a consequence of the physical association between H-Y antigen and &-microglobulin-H-2 antigen dimers on the male mouse plasma membrane. Cell lineage-specific antigens, on the other hand, have no reason to be involved in competitive displacements and, therefore, no reason to associate with P,-microglobulin-H-2 antigen dimers. Presently observed H-2-unrestricted killings of testicular Sertoli cells by female T cells might have been directed against such cell lineage-specific antigens. Experimental
Procedures
Animals In the absence of the Y chromosome, an autosomal dominant Sxr gene of the mouse induces testicular organization in XX embryos (Cattanach, Pollard and Hawkes, 1971), and these sex-reversed XX, Sxr/males express H-Y antigen in the absence of the Y (Bennett et al., 1977). These sterile males served as the source of pure Sertoli, as well as Leydig cells. Our stock of mice in which Sxr gene is maintained together with two X-lmked coat marker genes, Ta and B/o, initially went through six generations of brother x sister matings and was originally H-2k homozygous. H2d and H-2b homozygous sublines were subsequently established. H-2d was introduced via BLAB/C, and H-2” via C57BL/6. Pure Sertoli Cells from Young XX, Sxrl- Testes in the testicular environment, primordial germ cells having more than one X chromosome are incapable of neonatal differentiation into definitive spermatogonia (Cattanach et al., 1971). Accordingly, prepubertal as well as postpubertal testes of young XX, Sxr/males (5-100 days of age) abound with Sertoli cell-only seminiferous tubules (Figure 4, left). 1.5-2.0 x IO6 pure Sertoli cells were obtained from a pair of such testes by the following procedure. Bundles of seminiferous tubules gently teased out of each testis first underwent the 0.5% collagenase (Worthington) treatment for IO min at 37” C. This denuded seminiferous tubules and yielded a suspension of free extratubular elements, mostly Leydig cells. The 0.5% trypsin (Gibco) treatment of those denuded seminiferous tubules yielded a free suspension of Sertoli cells. When cultured, those Sertoli cells from prepubertal testes divided once or twice within the first 3 days while forming a monolayer (Figure 5, left). As XX, Sxrl- male mice grew older, seminiferous tubules progressively collapsed, and hypertrophic Leydig cells became the dominant cell type of the testis (Figure 4, right). The above noted collagenase treatment of older testes (110 days or more) enabled us to collect 3-5 x 1Q6 free Leydig cells per mouse. In Viva Immunization of H-2-Compatible Females with Sertoli Cells H-2d and H-2” homozygous females of our stock were immunized intraperitoneally with two or three doses of 1.5-2.0 x lo6 XX, Sxr/- Sertoli cells of the same H-2 type 8-9 days apart. Since a pair of testes barely sufficed to supply the above number of Sertoli cells, the availability of XX, Sxr/males became the limiting factor of our experiments. All together, nearly 50 males of the above genotypes were used. One week after the last immunization, females were killed and the spleen from each one of these fe.males was used as the source of presensitized T cells that were to be primed once more in vitro. Because we found that only females immunized 3 times yielded anti-Sertoli antibodies of sufficient strength, we discarded sera of twice-immunized females.
Cdl
648
Figure
4. Histological
Sections
of Testes
from
XX, Sxrl-Male
Mice
Lenses used: 45 x 10. (Left) from a 101 days old mouse. Sertoli cell-only cells. (Right) from a 182 days old mouse. Due to collapsing seminiferous source of Leydig cells.
seminiferous tubules,
tubules are still prominent. A good source of Sertoli hypertrophic Leydig cells dominate the field. A good
Further Sensitization in Vitro of T Cells Derived from Immunized Females From a spleen cell suspension of an immunized female, contaminating erythrocytes were first removed by the Tris-ammonium chloride (8.3 g/l) treatment, and then the nylon wool column filtration method was used to obtain a fraction enriched forT cells (Julius, Simpson and Herzenberg, 1973). By the above procedure, each female yielded IO-20 x IO6 T cells. They were then suspended together with 2 x lo6 freshly prepared Sertoli cells of the same H-2 type in 10 ml of medium RPM1 1640 containing 10% lgfree fetal calf serum (Grand Island), 10 mM HEPES, 5 x 10m5 M 2mercaptoethanol, penicillin and streptomycin, and cultured for 5 days in a humidified 10% COzatmosphere at 37°C. Because of the large size of Sertoli cells, the target-to-responder T cell ratio of 1:5 to I:1 0, instead of the customary 1 :I, sufficed for the in vitro sensitization. The T cell preparation for each experiment was derived from one immunized female; pooled preparations were never used. When XY male spleen cells, instead of XX, Sxrltesticular Sertoli ceils, were used to sensitize female T cells, we followed the protocol described by Gordon et al. (1976) in every detail. H2”/H-2” females were immunized with male spleen cells of the same heterozygous H-2 type.
however, all subsequent experiments were performed only at the 4:1 ratio. An appropriate concentration of these doubly primed T cells were partitioned in triplicate to polypropylene microtesttubes of 500 ~1 capacity (Bio-Rad). To each, freshly prepared Sertoli cells were added as targets to the volume of 250 ~1. The mixture was then incubated for 2.5 hr in the culture conditions already specified. Two controls were used: the first consisted of Sertoli cells alone, while in the second, unprimed virgin T cells obtained from the spleens of young nonimmunized females were used as attackers. In calculating a corrected percentage lysis, we used the second control. Preparations of Ovarian Follicular and Luteal Cells as Targets A suspension of free follicular (granulosa) cells was obtained from ovaries of 15-18 day old juvenile BALBlc females. Mediumsized growing follicles were collected under the dissecting microscope. They then underwent two successive enzyme treatments of 15 min each at 37”C-first with 0.5% trypsin and then with 0.5% hyaluronidase (Sigma). A pair of ovaries usually yielded 5 x lo5 free follicular cells. As the source of luteal cells, ovaries from 6-8 month old BALB/c females each containng several large corpora lutea were used. A suspension of free luteal cells was obtained by the double enzyme treatments described above.
Lysis by Doubly Primed Female T Cells of Target Sertoli Cells During 5 days of culture, surviving Sertoli cells formed a monolayer at the culture tube bottom. This enabled us to collect doubly primed T cells by a simple pipetting procedure. These collected T cells were washed once in a balanced salt solution, and then resuspended in Eagle’s minimal essential medium containing 10% Ig-free calf serum and 10 mM HEPES. Initially, attacker-totarget ratios between 8:1 and 2:1 were tried. When we found that the maximal lysis of targets already occurred at the 3:1 ratio,
Lysis of Target Sertoli Cells by Cytotoxic Female T Cells Because of their great size and nuclear morphology (Figure 5), target Sertoli cells were unmistakably distinct from tiny attacking T cells. Accordingly, we were able to witness the actual killings of Sertoli cells by attacking T cells under the microscope. Furthermore, the trypan blue dye exclusion test routinely used for the cytotoxicity assay of humoral antibodies was readily applicable to T cell killings of target Sertoli cells. A corrected percentage lysis of the targets was determined by the following formula:
Membrane 649
Figure
Antigens
5. Cultured
of Testicular
Sertoli
Sertoli
Cells Grown
Cells
as a Monolayer
(Left) untreated. The characteristic morphology of a large round Sertoli cell nucleus should be noted: one prominent nucleolus attached by one long and one short heterochromatic body. (Right) under the T ceil attack. Sertoli cells are being lysed by attached tiny T cells. Although there are crowds of T cells, actual killing is done by one or two attackers per target. Attackers and targets were H-2” homozygous. corrected
% lysis =
,OO x experimental
References dead
cell % - control
100 - control * Control
dead cell % = killing
dead
by unprimed
dead
cell* %
cell % T cells.
Since H-Y antigen is probably the cause of H-P-restricted killings of Sertoli cells, we expected such cytotoxic female T ceils to lyse male, but not female spleen cells of the same H-2 type. Conversely, when we encountered H-2-nonrestricted Sertoli cell killings, we wanted to determine whether this type of T cell also lyses ovarian follicular cells but not male spleen cells. Lysis of these other cell types had to be measured by the 5’Cr release test (Gordon, Simpson and Samelson, 1975). Unfortunately, the corrected percentage lyses of non-Sertoli targets by doubly primed female T cells that we obtained by the application of 5’Cr release tests were usually below 10% for the positive results, the highest being 13.3%. Aside from the use of H-Z” and H-2d haplotypes for most of our experiments to which the H-Y antigen-nonresponder status has been assigned (Simpson and Gordon, 1977), our to inexperience with the 5’Cr release test no doubt contributed those rather inconclusive results. When our H-2” subline of the Sxr stock is sufficiently expanded, we shall repeat these experiments with non-Sertoli target cells. Our inconclusive results, however, support the expectations noted above. Acknowledgments work was supported in part by an NIH contract grant. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. December
27,1977;
revised
February
2,1978
B. J., Scheid, M., Wachtel, B. M. (1977). Serological sex reversed phenotypic
Beutler, B., Nagai, Y., Ohno, S., Klein, G. and Shapiro, M. (1978). The HLA-dependent expression of testis-organizing H-Y antigen by human male cells. Cell 73, 509-513. Blank, K. J. and Lilly, F. (1977). Evidence for an H-2/viral protein complex on the cell surface as the basis of H-2 restriction of cytotoxicity. Nature 269, 808-809. Cattanach, 6. M., Pollard, C. E. and Hawkes, reversed mice: XX and X0 males. Cytogenetics
S. G. (1971). IO, 318-327.
Sex
Gordon, R. D., Simpson, E. and Samelson, L. E. (1975). In vitro cell mediated immune response to the male specific (H-Y) antigen in mice. J. Exp. Med. 142, 1108-1120. Gordon, R. D., Mathieson, B. J., Samelson, L. E., Boyse, E. A. and Simpson, E. (1978). The effect of allogeneic presensitization on H-Y graft survival and in vitro cell-mediated responses to H-Y antigen. J. Exp. Med. 744, 810-821. Gordon, R. D., Samelson, L. E. and Simpson, E. (1977). Selective response to H-Y antigen by F, female mice sensitized to F, male cells. J. Exp. Med. 746, 606-610. Krco, C. J. and Goldberg, detection on 8-cell embryos.
This
Received
Bennett, D., Boyse, E. A., Mathieson, S. S., Yanagisawa, W. and Cattanach, evidence for H-Y antigen in Sxr-XX males. Nature 265, 255-257.
E. A. (1976). H-Y (male) Science 793, 1134-1135.
Julius, M. H., Simpson, E. and Herzenberg, L. A. (1973). method for the isolation of functional thymus-derived lymphocytes. Eur. J. Immunol. 3, 645-649. McLaren.A. New York: Ohno,
(1976). Mammalian Cambridge University
S. (1977a).
The Y-linked
Chimaeras. Press). H-Y antigen
(Cambridge, locus
antigen: A rapid murine London,
and the X-linked
Cell 650
Tfm locus determining
as major regulatory genes of the mammalian mechanism. J. Steroid Blochem. 8, 585-592.
Ohno, S. (1977b). plasma membrane proteins. Immunol.
The original function anchorage site of Rev. 33, 59-69.
Ohno, S., Christian, L. C., Wachtel, Hormone-like role of H-Y antigen Nature 261, 597-599.
sex
of MHC antigens as the organogenesis-directing
S. S. and Koo, G. C. (1976). in bovine freemartin gonads.
Ohno, S., Nagai, Y. and Ciccarese, S. (1977). Testicular lysostripped of H-Y antigen organize ovarian follicle-like gates. Cytogenet. Cell Genet., in press.
cells aggre-
Ohno, S., Ciccarese, S., Nagai, Y. and Wachtel, S. S. (1978). H-Y antigen in testes of XX (BALB)/XY (C3H) chimaeric male mice. Arch. Androl. 1, 103-109. Peters, H. (1970). Migration gonad and their differentiation. 259,91-98.
of gonocytes into the mammalian Phil. Trans. Roy. Sot. Lond. B
Scheid, M., Boyse, E. A., Carswell, E. A. and Old, L. J. (1972). Serologically demonstrable alloantigens of mouse epidermal cells. J. Exp. Med. 735, 938-955. Simpson, E. and Gordon, R. D. (1977). Responsiveness to H-Y antigen Ir gene complementation and target cell specificity. Immunol. Rev. 35, 59-75. Wachtel, S. determination.
S. (1977). H-Y antigen Science 198, 797-799.
and
the
genetics
of
sex
Wachtel, S. S., Koo, G. C. and Boyse, E. A. (1975a). Evolutionary conservation of H-Y (male) antigen. Nature 254, 270-272. Wachtel, S. S., Ohno, S., Koo, G. C. and Boyse, E. A. (1975b). Possible role of H-Y antigen in primary sex determination. Nature 257,235-236. Zenzes, M. T., Wolf, U., Gunter, on the function of H-Y antigen: experiments of rat gonadal cells.
E. and Engel, W. (1977). Studies dissociation and reorganization Cytogenet. Cell Genet., in press.
Vol. 13, 643-650,
April
1978,
Copyright
0 1978 by MIT
Two Plasma Membrane Antigens of Testicular Sertoli Cells and H-2-Restricted versus Unrestricted Lysis by Female T Cells Salvatrice Ciccarese and Susumu Ohno Division of Biology City of Hope National Medical Center Duarte, California 91010
Summary Sertoli cell-only seminiferous tubules of sterile XX, Sxr/-male mice served as an excellent source of pure Sertoli cells. When H-2-compatible female mice were immunized 3 times with these Sertoli cells, resulting antibodies recognized two antigens on the plasma membrane of testicular Sertoli cells. They were male-specific, but ubiquitously expressed H-Y antigen and the cell lineage-specific antigen which Sertoli cells shared with ovarian follicular cells. Doubly primed (2 or 3 times in vivo, and once in vitro) cytotoxic T cells from these females lysed target Sertoli cells in both H-2-restricted and nonrestricted manners. While H-2-restricted killings were attributable to H-Y antigen, further work is needed to identify the Sertoli follicular cell lineage-specific antigen as the cause of H-2-nonrestricted killings. Introduction A testis-organizing role has been assigned to the male-specific but ubiquitously expressed H-Y antigen (Wachtel et al., 1975b; Ohno, 1977; Wachtel, 1977). Inasmuch as testicular organogenesis involves not one but several cell types, however, it was felt that the possession of H-Y antigen by all cell types-while an absolute prerequisite-is by itself insufficient for testicular organogenesis. A set of additional plasma membrane antigens, each specific for a particular cell type, appeared essential since somatic elements within the central gonadal bastema of an emerging male embryonic gonad must sort themselves out before precursors of Sertoli ceils can get together to form the wall of testicular cords while not involving precursors of Leydig ceils in this process. Such sorting out processes are likely to depend upon the cell typespecific plasma membrane antigen through which those belonging to the same lineage can recognize each other. In vitro reaggregation experiments using mouse and rat gonadal cells have shown that testicular Sertoli cells lysostripped of H-Y antigen behaved as though they were follicular cells (Ohno, Nagai and Ciccarese, 1977; Zenzes et al., 1977). This indicated that testicular Sertoli cells and ovarian follicular (granulosa) cells belong to the same lineage; thus they might be marked by the identical plasma membrane antigen. One outstanding characteristic of H-Y antigen is
that cytotoxic female T cells treat it exactly in the same manner as a viral antigen, so that lysis by T ceils of H-Y-incompatible target cells is as H-2restricted in the mouse as lysis of virally infected or transformed target cells (Gordon, Simpson and Samelson, 1975; Gordon et al., 1976; Simpson and Gordon, 1977). As to the cause of H-2 restriction, direct evidence of physical association between HY antigen and &-microglobulin-HLA antigen dimers on the plasma membrane of male human cells has recently been obtained (Beutler et al., 1978), and equally direct evidence of physical association between a Friend leukemia viral antigen and /3,-microglobulin-H-2 antigen dimers on the infected mouse plasma membrane has been published (Blank and Lilly, 1977). One of us proposed that the original function of p,-microglobulin-MHC antigen dimers has been to serve as the plasma membrane anchorage sites of organogenesis-directing antigens such as testisorganizing H-Y antigen, and that for this reason, parasitic viruses evolved the strategy of displacing these functionally important host antigens with their own (Ohno, 197713). According to this proposal, plasma membrane antigens that are not organogenesis-directing are not in association with p,-microglobulin-MHC antigen dimers. We therefore asked, whether female T cell killings of testicular Sertoli cells are invariably H-2-restricted. Results Cytotoxic Specificities of Anti-Sertoli Antibodies Serum samples from H-2” as well as H-2” females immunized 3 times with H-Pcompatible XX, Sxrltesticular Sertoli cells were invariably a mixture of two antibodies: H-Y antibody and another antibody which detected the cell lineage-specific antigen shared by testicular Sertoli cells and ovarian follicular cells. The above was deduced by combinations of cytotoxicity and absorption tests as shown in Table 1. In the presence of agarose-absorbed guinea pig complements, unabsorbed antibodies were cytotoxic to male but not female epidermal cells (BALB/c), to ovarian follicular but not luteal cells (BALB/c), and to testicular Sertoli but not primordial germ cells (Sxr stock). A slight cytotoxicity to Leydig cells was apparently due to H-Y antibody. Although most cell types of the mammalian male readily absorb H-Y antibody raised in the mouse (Wachtel, Koo and Boyse, 1975a), this antibody is cytotoxic only to very limited cell types of male mice; epidermal cells are one such type (Scheid et al., 1972). When raw anti-Sertoli antibodies were a,bsorbed with an excess number of either male epidermal cells (7 x lo6 cells per 50 ~1 of ‘14 diluted antibod-
Cell 644
Table 1. Cytotoxic Specificities Sertoli-Follicular (S-F) Antigen
of Anti-Sertoli
Sera That Contained
H-Y Antibody
Testicular Primordial Germ
and Antibody
Directed
against
a Cell Lineage-Specific
Ovarian
Sertoli
Leydig
Epidermal
Follicular
Luteal
Male
Female
+++
-
-
+++
t
++
-
H-Y antibody removed by absorption with male BALB spleen or epidermal cells -
+++
-
++
-
-
-
-
-
-
Unabsorbed
Further cells
absorption
with
ovarian
(++) 21-27% corrected percentage corrected percentage killing.
follicular -
killing
(see
Figures
1 and
2); (++)
ies) or male spleen cells (12 x 10” cells per the same amount), male-specific cytotoxicity toward epidermal cells, as well as a slight cytotoxicity toward testicular Leydig cells, was totally abolished (Table l), while absorption with female epidermal cells had no effect on their cytotoxicity toward male epidermal cells (Figure 1). Once H-Y antibody was removed by absorption with the above noted male cells, what remained of the anti-Sertoli antibodies became highly specific; among testicular elements, they were cytotoxic only to Sertoli cells of all ages (5-100 postnatal days), while XY neonatal primordial germ cells and XX, Sxrl-. Leydig cells remained totally unaffected. Anti-Sertoli antibodies deprived of H-Y antibody were also cytotoxic to BALB/c ovarian follicular cells, but not at all to luteal cells (Table 1). Although antibodies were consistently more cytotoxic to testicular Sertoli cells than to ovarian follicular cells (Figure 2), we believe that the difference was attributable to a relative impurity of the ovarian follicular cell fraction. Compared with the uniform appearance of Sertoli cell preparations under the microscope, our ovarian follicular cell fractions contained considerable proportions of spindle-shaped cells and rectangular cells that were not killed by anti-Sertoli antibodies. The indication that the plasma membrane of testicular Sertoli cells and ovarian follicular cells possesses an identical or very similar antigenic determinant was further confirmed by the absorption test. When H-Y antibody-removed anti-Sertoli antibodies were absorbed with ovarian follicular cells at the level of 12 x lo6 cells per 50 PI of l/4 diluted antibodies, their cytotoxicity toward testicular Sertoli cells was completely absolished (Figure 2). In fact, doubly absorbed (first with male spleen or epidermal cells, and then with ovarian follicular cells) anti-Sertoli sera demonstrated no residual cytotoxicity to any of the cell types listed in Table 1. Thus the immunization of H-2-compatible female mice with XX,
15-20%
corrected
percentage
killing
(see Figure
l-5%
H-Zd Q Thrice Immunized With H-2d Sertoli Cells
1
ABSORSEO 9 EPIOERMQ? \ x
"NPSSORSEOX esSORsE0
2); (k)
-7
0
.
C~EPIOERMQL \
x .
\
0 x
i
----b CLmpieme”tP
.
b Cootro,
----_____--------_
4 ANTIFigure BALB ments
SERTOLI
6 CELL
1. Cytotoxicity of Raw Anti-Sertoli Epidermal Cells in the Presence
16 SERA
3’2 DILUTION
Antibodies of Guinea
toward Male Pig Comple-
This serum sample was absorbed with male and female BALB epidermal cells at the level of 2.5 x lo6 cells per 50 ~1 ‘/a diluted antibodies. Only male epidermal cells and male spleen cells removed the cytotoxicity. This revealed the presence of H-Y antibody within raw anti-Sertoli antibodies.
sxrl-testicular Sertoli cells apparently provoked the production of two antibodies: one directed against testis-organizing H-Y antigen, and the other directed against the cell lineage-specific antigen which marks testicular Sertoli cells and ovarian follicular cells as being derived from the same cell lineage. This is not to suggest, of course, that the above are the only antigenic determinants on the plasma membrane of testicular Sertoli cells. Indeed, xeno-
Membrane 545
Antigens
of Testicular
Sertoli
Ceils
H-Zd Q Thrice Immunized With
ANTI -SERTOLI
CELLS
Figure 2. Reaction of Anti-Sertoli Ovarian Follicular Cells
(-anti-HY) Antibody
SERA with
DILUTION
Sertoli
Cells and
Once H-Y antibody was removed by absorption, the remaining anti-Sertoli antibody was cytotoxic only to testicular Sertoli cells themselves and to ovarian follicular cells. The absorption by ovarian follicular cells at the level of 12 x lo6 cells per 50 PI ‘/4 diluted antibody completely removed its cytoxicity toward Sertoli cells. Thus it became clear that this antibody was detecting an identical or very similar antigenic determinant on the plasma membrane of testicular Sertoli cells and ovarian follicular cells.
genie antibodies such as rabbit anti-mouse Sertoli antibodies appear to recognize a great variety of antigens that are difficult to sort out. At any rate, the two presently identified appear most pertinent to testicular organogenesis. Needless to say, antiSertoli-follicular cell lineage-specific antibody was an autoantibody as far as immunized females were concerned. We obviously need to do an histological, as well as functional study of the ovaries of immunized females. Our impression, however, is that this cell lineage-specific antigen on the plasma membrane of follicular cells rapidly disappears after puberty, and it is for this reason that adult H2-compatible females were able to generate an antibody directed against this antigenic determinant. Whether the corresponding antigenic determinant found on XX, Sxr/- testicular sertoli cells can also be found on adult XY Sertoli cells has not been determined either. It is possible that in the absence of male germ cells, XX, Sxrl- Sertoli ceils can never attain the differentiated state characteristically postpubertal. Nevertheless, anti-Sertoli sera were cytotoxic to a crude fraction of juvenile XY Sertoli cells.
H-2 Restricted and Unrestricted Cells by Doubly Primed Female
Lysis of Sertoli T Cells
A highlight of this preliminary study on T cell killings of Sertoli cells was our direct observation of the actual killings under the microscope (Figure 5, right). Surprisingly, unprimed virgin female T cells of the control mixture readily attached themselves to the plasma membrane of Sertoli cells. Such attachments, however, caused no death of Sertoli cells, even when the large cell surface of a single Sertoli cell was completely covered by 40 or more small lymphocytes. In sharp contrast, doubly primed female T cells were extremely efficient killers in the H-2-restricted as well as unrestricted situation. Lysis of a Sertoli cell was often caused by the attachment of only one or two tiny T cells. Under the attack, a large Sertoli cell cytoplasm literally boiled as it became increasingly vacuolated. Then the nucleus suddenly became pale and swollen. An individual killing took no more than one hour. Even when there were 20 or more T cells in the close vicinity of one Sertoli cell, the actual killing appeared to have been done only by one or two T cells. The observed efficacy of cytotoxic T cells apparently was a primary reason for the fact that in spite of a great difference in their sizes, the maximal killing was already attained at the attacker-to-target ratio of 3:l (Figure 3). Female mice primed in vivo as well as in vitro with H-2-compatible male XY spleen cells generate cytotoxic T cells directed against H-Y antigen and lyse male target cells in the H-2-restricted manner (Gordon et al., 1975, 1976; Simpson and Gordon, 1977). One might expect such primed female T cells to recognize H-Y antigen on the plasma membrane of Sertoli cells as well, and to lyse them also in the H-2-restricted manner. Unfortunately, the above investigators have also shown that H-2” and H-2k females are least efficient in generating cellmediated immune responses against H-2-compatible male target cells. Most of the mated pairs of our Sxr stock were either H-2k or H-P homozygous; the H-2b subline just begun constituted a distinct minority. Thus we chose to immunize BALE/B6 (H2”lH-2”) F, females with male F, spleen cells in the hope that cytotoxic T cells that readily lyse not only H-2” male spleen cells, but also H-2” testicular Sertoli cells, could be generated. Gordon, Samelson and Simpson (1977), however, have shown that BALB/BlO (H-2”/H-2b) F, female T cells sensitized with male F, spleen cells readily lysed F, as well as those H-2b male spleen cells presented as their targets, but did not lyse H-2” male spleen cells. Accordingly, it is quite possible that had we presented H-2” male spleen cells as targets to our doubly primed H-2”/H-2b female cytotoxic T cells,
Cell 646
TARGET SERTOLl H-2d H-2k
ATT. T-CELLS H-2d/H-2b PRIMED IN VIVO-IN VITRO 0-O o-o UNPRIMED x-x
Table 2. H-2 Restricted Testicular Sertoli Cells
and Unrestricted by H-2 Compatibly Target
H-2d
l -------. \
Primed GSpleen
with:
Responder ? T Cells
Lysis of XX, Sxr/Primed Female T Cells Testicular
H-2 Type
Sertoli
Cells
H-2 Restriction
Lysis
Cells +++ - Yes
\
J
.
GSpleen
Cells ++ - Yes
H-2d/H-2b
“1 -__-x--------o------------~
Sertoli
Sertoli
\
Sertoli
ATTACKERS
3:l
: TARGETS
2:l
0:l
RATIOS
Figure 3. H-P-Restricted Lysis of XX, Sxr/-Testicular by H-2d/H-2b (BALB/BG) F, Female T Cells Doubly Male F, H-2d/H-2b Spleen Cells
Sertoli Primed
Cells with
H-2d but not H-2k targets were lysed. Target nuclei stained by trypan blue were scored as lysed cells. The attacker-to-target ratio of 3:l already attained the near-maximal killing of H-2-compatible targets. At the 4:l ratio, the experimental killing of H-Y targets was 47.5%, whereas the corresponding value by unprimed virgin H-p/H-2” female T cells was 23.8%. Thus the corrected percentage lysis of H-T Sertoli cells was calculated to be 31 .I%, whereas the corrected percentage lysis of H-2” Sertoli cells at the 4:l ratio was 0.8%. Thus in Table 2, The first value was scored as (+++), and the second value as (-). 250 ~1 of a reaction mixture always contained 1.5 x lo5 Sertoli cells as targets to which an appropriate number of doubly primed female T cells was added to the desired attacker-to-target ratio.
they might not have lysed them. Nevertheless, when H-2” testicular Sertoli cells were presented as the targets, they readily lysed them, while they were not at all cytotoxic to H-2k Sertoli cells (Figure 3 and Table 2). Thus it appeared that if directed against H-Y antigen, female T cell killing of testicular Sertoli cells is H-2-restricted. It should be pointed out that H-Y antigen sites are apparently far more prominently displayed on the plasma membrane of Sertoli cells than on that of male spleen cells. For the routine production of H-Y antibody, we give five weekly injections of 525 x IO6 male spleen cells each, yet only 25% of the immunized females yield H-Y antibody of usable strengths. In sharp contrast, only three consecutive injections of 1.5-2.0 x lo6 XX, SxrlSertoli cells each generated usable H-Y antibody in every instance as already noted. The difference, however, was confined to the frequency of success. When H-2” and H-2” females were doubly primed
H-2” H-T
H-2k
H-2” H-2d H-2k/H-2” H-2b
+++:
Yes
++t ++;
No +
Cells
H-2“
Sertoli
H-2* Cells
x------x
I
4:1
Cells
H-2*
H-2d H-2k H-2d/H-2k
++ ++ ++
No
H-F
H-P H-2b
:I
No
Cells
H-2” Attacker-to-target age lysis; (+++) 3); (++) 21-30% rected percentage
ratio (4:l). (++++) 41-46% corrected percent31-40% corrected percentage lysis (see Figure corrected percentage lysis; (+) il-20% corlysis; (3) 5-10% corrected percentage lysis.
(2 or 3 times in vivo, and once in vitro) not with male spleen cells but with testicular Sertoli cells, lyses by sensitized female T cells of target Sertoli cells were seldom H-2-restricted (Table 2). It is probable that on the plasma membrane of Sertoli cells, H-2-compatible female T cells recognized not only H-Y antigen, but also other antigenic determinants such as the Sertoli-follicular cell lineagespecific antigen which their B cells recognized, and that their cytotoxic effect upon target Sertoli cells, if directed against the latter antigen, was not H-2-restricted. Whether or not sensitizing Sertoli cells generate two or more kinds of cytotoxic female T cells can be tested by offering these T cells male spleen cells and ovarian follicular cells as targets. Lyses of these target cell types, however, have to be measured by the 51Cr release test. As already noted, we have thus far obtained unsatisfactorily low percentage lyses by this method. Nevertheless, if these Sertoli cell-sensitized female T cells lysed male but not female spleen cells in the H-2-restricted manner, there existed T cells directed against H-Y antigen. If these female T cells lysed ovarian follicular cells, this time in H-Bunrestricted fashion, cytotoxic T cells directed against the Sertoli-follicular cell lineage-specific antigen
Membrane 647
Antigens
were obviously killings of Sertoli
of Testicular
responsible cells.
Sertoli
for
Cells
H-Punrestricted
Discussion Although male-specific H-Y antigen is ubiquitously expressed by nearly all the cell types of mammalian males. At first glance, the assignment of a specific organogenesis role to an antigen so ubiquitously expressed might seem rather odd. Yet to direct organogenesis, the plasma membrane component responsible for it has to express itself before the commencement of that organogenesis. Indeed, in the development of mouse embryos, gonadal differentiation does not begin until the twelfth day (Peters, 1970). H-Y antigen, however, is first seen 10 days earlier in preimplantation male embryos of the 8 cell stage (Krco and Goldberg, 1976). The ubiquitous expression of H-Y antigen may be viewed as the natural consequence of a need to express itself so early in embryonic development. Nevertheless, for H-Y antigen (uniformly posessed by all the cell types of the male gonad) to direct testicular organization, it has to depend upon a set of cell type-specific antigens that enabie individual cells of the same lineage to identify each other. One such antigen found on the plasma membrane of Sertoli cells was also identified on ovarian follicular cells. Such sex-independent expression of the cell lineage-specific plasma membrane antigen by gonadal cells might be the very factor that enables H-Y antigen to play a pivotal role in the primary (gonadal) sex determination of mammals. Indeed, in vitro reaggregation experiments have shown that if lysostripped of their H-Y antigen, testicular Sertoli cells readily behaved as though they were ovarian follicular cells (Ohno et al., 1977; Zenzes et al., 1977). XX/XY gonads of experimentally produced blastocyst fusion chimaeras most often organize testes, less often ovaries and seldom ovotestes (McLaren, 1976). This suggests the presence of a development strategy that avoids the formation of mosaic organs. It has been suggested that this strategy is based on competitive displacements on the plasma membrane of one organogenesis-directing antigen with another, and that.such competitive displacements are made possible because competing antigens, such as testis-organizing H-Y antigen versus a yet to be identified ovary-organizing antigen, utilize the same plasma membrane anchorage site which is furnished by P,-microglobulin-MHC antigen dimers (Ohno, 1977b). Our previous finding on bovine freemartin gonads indicated a short-range hormone-like role of H-Y antigen (Ohno et al., 1976). Furthermore, XX cells that comprised 50% of the Sertoli, Leydig cell population in XX/XY chimaeric mouse gonads were found to be en-
dowed with as much H-Y antigen as XY cells (Ohno et al., 1978). The fact that female T cell killing of male target cells, if directed against H-Y antigen, is H-2-restricted may be viewed as a consequence of the physical association between H-Y antigen and &-microglobulin-H-2 antigen dimers on the male mouse plasma membrane. Cell lineage-specific antigens, on the other hand, have no reason to be involved in competitive displacements and, therefore, no reason to associate with P,-microglobulin-H-2 antigen dimers. Presently observed H-2-unrestricted killings of testicular Sertoli cells by female T cells might have been directed against such cell lineage-specific antigens. Experimental
Procedures
Animals In the absence of the Y chromosome, an autosomal dominant Sxr gene of the mouse induces testicular organization in XX embryos (Cattanach, Pollard and Hawkes, 1971), and these sex-reversed XX, Sxr/males express H-Y antigen in the absence of the Y (Bennett et al., 1977). These sterile males served as the source of pure Sertoli, as well as Leydig cells. Our stock of mice in which Sxr gene is maintained together with two X-lmked coat marker genes, Ta and B/o, initially went through six generations of brother x sister matings and was originally H-2k homozygous. H2d and H-2b homozygous sublines were subsequently established. H-2d was introduced via BLAB/C, and H-2” via C57BL/6. Pure Sertoli Cells from Young XX, Sxrl- Testes in the testicular environment, primordial germ cells having more than one X chromosome are incapable of neonatal differentiation into definitive spermatogonia (Cattanach et al., 1971). Accordingly, prepubertal as well as postpubertal testes of young XX, Sxr/males (5-100 days of age) abound with Sertoli cell-only seminiferous tubules (Figure 4, left). 1.5-2.0 x IO6 pure Sertoli cells were obtained from a pair of such testes by the following procedure. Bundles of seminiferous tubules gently teased out of each testis first underwent the 0.5% collagenase (Worthington) treatment for IO min at 37” C. This denuded seminiferous tubules and yielded a suspension of free extratubular elements, mostly Leydig cells. The 0.5% trypsin (Gibco) treatment of those denuded seminiferous tubules yielded a free suspension of Sertoli cells. When cultured, those Sertoli cells from prepubertal testes divided once or twice within the first 3 days while forming a monolayer (Figure 5, left). As XX, Sxrl- male mice grew older, seminiferous tubules progressively collapsed, and hypertrophic Leydig cells became the dominant cell type of the testis (Figure 4, right). The above noted collagenase treatment of older testes (110 days or more) enabled us to collect 3-5 x 1Q6 free Leydig cells per mouse. In Viva Immunization of H-2-Compatible Females with Sertoli Cells H-2d and H-2” homozygous females of our stock were immunized intraperitoneally with two or three doses of 1.5-2.0 x lo6 XX, Sxr/- Sertoli cells of the same H-2 type 8-9 days apart. Since a pair of testes barely sufficed to supply the above number of Sertoli cells, the availability of XX, Sxr/males became the limiting factor of our experiments. All together, nearly 50 males of the above genotypes were used. One week after the last immunization, females were killed and the spleen from each one of these fe.males was used as the source of presensitized T cells that were to be primed once more in vitro. Because we found that only females immunized 3 times yielded anti-Sertoli antibodies of sufficient strength, we discarded sera of twice-immunized females.
Cdl
648
Figure
4. Histological
Sections
of Testes
from
XX, Sxrl-Male
Mice
Lenses used: 45 x 10. (Left) from a 101 days old mouse. Sertoli cell-only cells. (Right) from a 182 days old mouse. Due to collapsing seminiferous source of Leydig cells.
seminiferous tubules,
tubules are still prominent. A good source of Sertoli hypertrophic Leydig cells dominate the field. A good
Further Sensitization in Vitro of T Cells Derived from Immunized Females From a spleen cell suspension of an immunized female, contaminating erythrocytes were first removed by the Tris-ammonium chloride (8.3 g/l) treatment, and then the nylon wool column filtration method was used to obtain a fraction enriched forT cells (Julius, Simpson and Herzenberg, 1973). By the above procedure, each female yielded IO-20 x IO6 T cells. They were then suspended together with 2 x lo6 freshly prepared Sertoli cells of the same H-2 type in 10 ml of medium RPM1 1640 containing 10% lgfree fetal calf serum (Grand Island), 10 mM HEPES, 5 x 10m5 M 2mercaptoethanol, penicillin and streptomycin, and cultured for 5 days in a humidified 10% COzatmosphere at 37°C. Because of the large size of Sertoli cells, the target-to-responder T cell ratio of 1:5 to I:1 0, instead of the customary 1 :I, sufficed for the in vitro sensitization. The T cell preparation for each experiment was derived from one immunized female; pooled preparations were never used. When XY male spleen cells, instead of XX, Sxrltesticular Sertoli ceils, were used to sensitize female T cells, we followed the protocol described by Gordon et al. (1976) in every detail. H2”/H-2” females were immunized with male spleen cells of the same heterozygous H-2 type.
however, all subsequent experiments were performed only at the 4:1 ratio. An appropriate concentration of these doubly primed T cells were partitioned in triplicate to polypropylene microtesttubes of 500 ~1 capacity (Bio-Rad). To each, freshly prepared Sertoli cells were added as targets to the volume of 250 ~1. The mixture was then incubated for 2.5 hr in the culture conditions already specified. Two controls were used: the first consisted of Sertoli cells alone, while in the second, unprimed virgin T cells obtained from the spleens of young nonimmunized females were used as attackers. In calculating a corrected percentage lysis, we used the second control. Preparations of Ovarian Follicular and Luteal Cells as Targets A suspension of free follicular (granulosa) cells was obtained from ovaries of 15-18 day old juvenile BALBlc females. Mediumsized growing follicles were collected under the dissecting microscope. They then underwent two successive enzyme treatments of 15 min each at 37”C-first with 0.5% trypsin and then with 0.5% hyaluronidase (Sigma). A pair of ovaries usually yielded 5 x lo5 free follicular cells. As the source of luteal cells, ovaries from 6-8 month old BALB/c females each containng several large corpora lutea were used. A suspension of free luteal cells was obtained by the double enzyme treatments described above.
Lysis by Doubly Primed Female T Cells of Target Sertoli Cells During 5 days of culture, surviving Sertoli cells formed a monolayer at the culture tube bottom. This enabled us to collect doubly primed T cells by a simple pipetting procedure. These collected T cells were washed once in a balanced salt solution, and then resuspended in Eagle’s minimal essential medium containing 10% Ig-free calf serum and 10 mM HEPES. Initially, attacker-totarget ratios between 8:1 and 2:1 were tried. When we found that the maximal lysis of targets already occurred at the 3:1 ratio,
Lysis of Target Sertoli Cells by Cytotoxic Female T Cells Because of their great size and nuclear morphology (Figure 5), target Sertoli cells were unmistakably distinct from tiny attacking T cells. Accordingly, we were able to witness the actual killings of Sertoli cells by attacking T cells under the microscope. Furthermore, the trypan blue dye exclusion test routinely used for the cytotoxicity assay of humoral antibodies was readily applicable to T cell killings of target Sertoli cells. A corrected percentage lysis of the targets was determined by the following formula:
Membrane 649
Figure
Antigens
5. Cultured
of Testicular
Sertoli
Sertoli
Cells Grown
Cells
as a Monolayer
(Left) untreated. The characteristic morphology of a large round Sertoli cell nucleus should be noted: one prominent nucleolus attached by one long and one short heterochromatic body. (Right) under the T ceil attack. Sertoli cells are being lysed by attached tiny T cells. Although there are crowds of T cells, actual killing is done by one or two attackers per target. Attackers and targets were H-2” homozygous. corrected
% lysis =
,OO x experimental
References dead
cell % - control
100 - control * Control
dead cell % = killing
dead
by unprimed
dead
cell* %
cell % T cells.
Since H-Y antigen is probably the cause of H-P-restricted killings of Sertoli cells, we expected such cytotoxic female T ceils to lyse male, but not female spleen cells of the same H-2 type. Conversely, when we encountered H-2-nonrestricted Sertoli cell killings, we wanted to determine whether this type of T cell also lyses ovarian follicular cells but not male spleen cells. Lysis of these other cell types had to be measured by the 5’Cr release test (Gordon, Simpson and Samelson, 1975). Unfortunately, the corrected percentage lyses of non-Sertoli targets by doubly primed female T cells that we obtained by the application of 5’Cr release tests were usually below 10% for the positive results, the highest being 13.3%. Aside from the use of H-Z” and H-2d haplotypes for most of our experiments to which the H-Y antigen-nonresponder status has been assigned (Simpson and Gordon, 1977), our to inexperience with the 5’Cr release test no doubt contributed those rather inconclusive results. When our H-2” subline of the Sxr stock is sufficiently expanded, we shall repeat these experiments with non-Sertoli target cells. Our inconclusive results, however, support the expectations noted above. Acknowledgments work was supported in part by an NIH contract grant. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. December
27,1977;
revised
February
2,1978
B. J., Scheid, M., Wachtel, B. M. (1977). Serological sex reversed phenotypic
Beutler, B., Nagai, Y., Ohno, S., Klein, G. and Shapiro, M. (1978). The HLA-dependent expression of testis-organizing H-Y antigen by human male cells. Cell 73, 509-513. Blank, K. J. and Lilly, F. (1977). Evidence for an H-2/viral protein complex on the cell surface as the basis of H-2 restriction of cytotoxicity. Nature 269, 808-809. Cattanach, 6. M., Pollard, C. E. and Hawkes, reversed mice: XX and X0 males. Cytogenetics
S. G. (1971). IO, 318-327.
Sex
Gordon, R. D., Simpson, E. and Samelson, L. E. (1975). In vitro cell mediated immune response to the male specific (H-Y) antigen in mice. J. Exp. Med. 142, 1108-1120. Gordon, R. D., Mathieson, B. J., Samelson, L. E., Boyse, E. A. and Simpson, E. (1978). The effect of allogeneic presensitization on H-Y graft survival and in vitro cell-mediated responses to H-Y antigen. J. Exp. Med. 744, 810-821. Gordon, R. D., Samelson, L. E. and Simpson, E. (1977). Selective response to H-Y antigen by F, female mice sensitized to F, male cells. J. Exp. Med. 746, 606-610. Krco, C. J. and Goldberg, detection on 8-cell embryos.
This
Received
Bennett, D., Boyse, E. A., Mathieson, S. S., Yanagisawa, W. and Cattanach, evidence for H-Y antigen in Sxr-XX males. Nature 265, 255-257.
E. A. (1976). H-Y (male) Science 793, 1134-1135.
Julius, M. H., Simpson, E. and Herzenberg, L. A. (1973). method for the isolation of functional thymus-derived lymphocytes. Eur. J. Immunol. 3, 645-649. McLaren.A. New York: Ohno,
(1976). Mammalian Cambridge University
S. (1977a).
The Y-linked
Chimaeras. Press). H-Y antigen
(Cambridge, locus
antigen: A rapid murine London,
and the X-linked
Cell 650
Tfm locus determining
as major regulatory genes of the mammalian mechanism. J. Steroid Blochem. 8, 585-592.
Ohno, S. (1977b). plasma membrane proteins. Immunol.
The original function anchorage site of Rev. 33, 59-69.
Ohno, S., Christian, L. C., Wachtel, Hormone-like role of H-Y antigen Nature 261, 597-599.
sex
of MHC antigens as the organogenesis-directing
S. S. and Koo, G. C. (1976). in bovine freemartin gonads.
Ohno, S., Nagai, Y. and Ciccarese, S. (1977). Testicular lysostripped of H-Y antigen organize ovarian follicle-like gates. Cytogenet. Cell Genet., in press.
cells aggre-
Ohno, S., Ciccarese, S., Nagai, Y. and Wachtel, S. S. (1978). H-Y antigen in testes of XX (BALB)/XY (C3H) chimaeric male mice. Arch. Androl. 1, 103-109. Peters, H. (1970). Migration gonad and their differentiation. 259,91-98.
of gonocytes into the mammalian Phil. Trans. Roy. Sot. Lond. B
Scheid, M., Boyse, E. A., Carswell, E. A. and Old, L. J. (1972). Serologically demonstrable alloantigens of mouse epidermal cells. J. Exp. Med. 735, 938-955. Simpson, E. and Gordon, R. D. (1977). Responsiveness to H-Y antigen Ir gene complementation and target cell specificity. Immunol. Rev. 35, 59-75. Wachtel, S. determination.
S. (1977). H-Y antigen Science 198, 797-799.
and
the
genetics
of
sex
Wachtel, S. S., Koo, G. C. and Boyse, E. A. (1975a). Evolutionary conservation of H-Y (male) antigen. Nature 254, 270-272. Wachtel, S. S., Ohno, S., Koo, G. C. and Boyse, E. A. (1975b). Possible role of H-Y antigen in primary sex determination. Nature 257,235-236. Zenzes, M. T., Wolf, U., Gunter, on the function of H-Y antigen: experiments of rat gonadal cells.
E. and Engel, W. (1977). Studies dissociation and reorganization Cytogenet. Cell Genet., in press.
