(CANCER RESEARCH 49. 1207-1213, March 1, 1989]

Specificity of Antigens on UV Radiation-induced Antigenic Tumor Cell VariantsMeasured in Vitro and in Vivo1

Lisa W. Hostetler,2 Cynthia A. Romerdahl,3 and Margaret L. Kripke4

Department of Immunology; The University of Texas M. D. Anderson Cancer Center, Houston, Texas 77030

ABSTRACT

In vitro exposure of nonimmunogenic murine tumor cells to UVradiation (UVR) generates highly antigenic variants that are immunolog-ically rejected by normal, syngeneic mice. The purpose of this study wasto determine whether these antigenic variants cross-react immunologi-cally with the parental tumor and whether the UVR-associated antigenunique to UVR-induced tumors is also present on the variants. Antigenic(régresser)variants and nonimmunogenic (progresser) clones derivedfrom UV-irradiated cultures of the C3H K1735 melanoma and SF19spontaneous fibrosarcoma cell lines were used to address these questions.In an in vivo immunization and challenge assay, the antigenic variantsdid not induce cross-protection among themselves, but each inducedimmunity against the immunizing variant, the parent tumor cells, andnonimmunogenic clones derived from UV-irradiated parent cultures.Therefore, the variants can be used to induce in mice a protectiveimmunity that prevents the growth of the parent tumor and nonimmunogenic clones, but not other antigenic variants. In contrast, immunizationwith cells of the parental tumor or the nonimmunogenic clones inducedno protective immunity against challenge with any of the cell lines.Utilizing the K1735 melanoma-derived cell lines in vitro, T-helper (Th)cells isolated from tumor-immunized mice were tested for cross-reactivityby their ability to collaborate with trinitrophenyl-primed B-cells in thepresence of trinitrophenyl-conjugated tumor cells. Also, the cross-reactivity of cytotoxic T-lymphocytes from tumor-immunized mice was assessed by a 4-h "Cr-release assay. Antigenic variants induced cytotoxicT-lymphocytes and Th activity that was higher than that induced by theparent tumor and nonimmunogenic clones from the UVR-exposed parenttumor and cross-reacted with the parental tumor cells and nonimmunogenic clones, but not with other antigenic variants. Furthermore, upontransplantation, the UVR-induced antigenic variants grew in UV-irradiated and immunosuppressed mice, but not in untreated mice indicatingthat the variants expressed the determinant recognized by suppressor I -cells present in UV-irradiated mice. These results demonstrate thathighly antigenic cells generated by the in vitro exposure of two differentmurine tumors to UV radiation express a determinant shared with theparental tumor cells and nonimmunogenic clones, a unique variant-specific determinant and the suppressor cell-defined determinant present onUVR-induced tumors. Based on these results, two models are proposedto explain the make-up of the antigenic determinants present on theUVR-induced antigenic variants.

INTRODUCTION

Chronic exposure of mice to (280-320 nm) UVR5 results in

the production of highly antigenic skin cancers that fail to growwhen transplanted into normal, syngeneic recipients (1). Therejection is associated with an immune response specific for the

Received 7/27/88; revised 11/23/88; accepted 11/30/88.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 inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

' This work was supported by the Mary Kay Ash Foundation.2 A University of Texas at Houston MD/PhD Program Fellow.3 Formerly a Tesoro Petroleum Corporation Postdoctoral fellow. Present

address: Department of Zoology, Oregon State University, Corvallis, OR 97331.' To whom requests for reprints should be addressed, at Department of

Immunology, P. O. Box 178, The University of Texas M. D. Anderson CancerCenter, 1515 Holcombe Boulevard, Houston, TX 77030.

5The abbreviations used are: UVR, ultraviolet radiation; Ts, suppressor T-cells; CTL, cytotoxic T-lymphocyte(s); Th, T-helper cell; NW, nylon wool; HRBC,horse red blood cell; PFC, plaque-forming cell; FBS, fetal bovine serum; PBS,phosphate-buffered saline; TNP, trinitrophenyl; HBSS, Hanks' buffered salinesolution; MNNG, /V-methyl-N'-nitrosoguanidine.

immunizing tumor, indicating the presence of antigens uniqueto each tumor (2, 3). In contrast, every UVR-induced tumor is

recognized by Ts present in the lymphoid organs of micechronically irradiated with UVR (2, 4, 5). Even though thesesame tumors express individually specific transplantation antigens, the Ts appear to inhibit the immune response to allcancers induced by UVR but not the response to other alloge-neic or syngeneic tumors (6). Recognition of UVR-inducedtumors as a group by UVR-induced Ts cells suggests that theyexpress a common regulatory antigen, in addition to a uniquedeterminant. Roberts demonstrated that a cloned suppressorcell line derived from spleen cells of UV-irradiated mice recognizes multiple tumors, providing direct evidence that a common determinant expressed on all UVR-induced tumors is thetarget for the Ts cells (7).

Antigens on UVR-induced tumors have also been demonstrated in vitro by utilizing CTL assays and an indirect assayfor Th activity. Individually specific antigens have been demonstrated on UVR-induced tumors by assessing the specificityof CTL (8, 9) and Th cells (10) generated after immunizationof mice with UVR-induced tumors. In addition, Schreiber andcolleagues (11, 12) have isolated and cloned CTL lines frommice immunized with UVR-induced tumors. These cloned CTLrecognize transplantation antigens specific for the immunizingtumor and show no evidence of cross-reactivity with othertumors whether or not they are UVR-induced.

Recently we reported that in vitro exposure of cells from twoweakly immunogenic murine tumors, a melanoma and a fibrosarcoma, to UVR generated a high frequency of antigenic(régresser)variants. Approximately 50% of the clones isolatedfrom UV-irradiated cultures of the melanoma and fibrosarcomawere immunologically rejected in normal, syngeneic mice butgrew progressively in immunosuppressed mice (13, 14). Antigenic variants derived from the fibrosarcoma cell lines were likeUVR-induced tumors in that the variants were not cross-protective in vivo, but were recognized by UVR-induced Ts (13).In this study, we investigated whether antigenic variants derivedfrom the UV-irradiated melanoma cells also possess the antigenic characteristics of UVR-induced tumors or whether theinduction of UVR-associated antigens was a unique propertyof the fibrosarcoma-derived variants. Second, we wished todetermine whether the antigenic variants share antigens withthe parental cells from which they are derived that can bedetected by //; vivo immunization and challenge. In addition,we compared the specificity of the variant-induced immuneresponse observed in vivo with that of Th and CTL obtainedfrom variant-immunized mice and assayed in vitro.

MATERIALS AND METHODSMice. Specific pathogen-free C3H/HeN (MTV") mice were supplied

by the Animal Production Area of the Frederick Cancer ResearchCenter. Mice were used at 8-12 weeks of age for all experiments, andwithin a single experiment, all mice were age- and sex-matched. Theanimals were housed in a pathogen-free, barrier facility where ambientlight was automatically controlled to produce 12-h light, 12-h darkcycles; they were maintained in a facility accredited by the American

1207

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIGENS ON UVR-INDUCED ANTIGENIC VARIANTS

Association for Accreditation of Laboratory Animal Care according tothe NIH guide for Laboratory Animal Care.

Tumors. K1735 is a melanoma cell line derived from a tumor inducedin a C3H/HeN (MTV") mouse by UVR and croton oil treatment (15).

SF19 fibrosarcoma is a spontaneous tumor that arose in a C3H/HeN(MTV") mouse (16). Both of these cell lines grow progressively when

transplanted into syngeneic mice. Tumor cell lines K1735-OR.5,K1735-LD.1, SF19-OR.26, and SF19-OR.58 are clones derived fromUV-irradiated K1735 or SF19 cultures following in vitro exposure toUVR as described previously (13, 14). These cell lines are highlyantigenic in that they are rejected following s.c. injection into normalsyngeneic mice. The K1735-OR.3 and SF19-LD.3 cell lines are non-immunogenic (progresser) clones derived from UV-irradiated K1735and SF19 cultures, respectively (13,14). UV 2240 was induced in C3H/HeN (MTV~) mice by chronic exposure to UVR. It grows in UV-

irradiated and immunosuppressed mice, but not untreated syngeneichosts (17). The MCA-FM1 cell line, produced by Dr. Philip Frost(Department of Cell Biology, University of Texas M. D. AndersonCancer Center, Houston, TX), is an antigenic (régresser)variant derivedfrom the 3-methylcholanthrene-induced MCA-F fibrosarcoma of C3H/HeJ strain origin after in vitro exposure to the mutagen 7V-methyl-A"-nitro-JV-nitrosoguanidine (MNNG). This cell line is immunologicallyrejected upon transplantation into normal syngeneic mice, and wasobtained from Dr. Stephen LeGrue (Department of Immunology, University of Texas M. D. Anderson Cancer Center, Houston, TX).

All tumors were maintained in tissue culture using Eagle's minimal

essential medium (GIBCO, Grand Rapids, MI) supplemented with 10%FBS, 2% vitamin solution, 10% sodium p>nivale. 1% nonessentialamino acids, and 1% L-glutamine, without antibiotics (complete minimal essential media). All cell lines were tested routinely to assure theabsence of Mycoplasma, and all lines were free of pathogenic murineviruses (Microbiological Associates, Rockville, MD). Subconfluentmonolayers of the tumor cells were harvested by a 30-s trypsinization(0.25% trypsin-versene). The single-cell suspensions were washed andresuspended in HBSS (GIBCO). Cell number was determined using ahemacytometer, and viability was measured by the exclusion of trypanblue.

In Vivo Cross-Protection Experiment. Normal C3H mice were immunized by s.c. injection of 2 x IO6 tumor cells on Day 0. Because

both progresser and régressertumors were used for immunization, allcell lines were exposed to 10,000-rad \ irradial ion before injection tomake the immunizing inocula comparable. Control, unimmunized micewere injected with an equal volume of culture medium. On Day 14,after regression of the immunizing tumors, the mice were exposed to450 rads of X-irradiation. This procedure reduces the primary responseagainst challenge implants while the secondary response in immunizedmice is left intact (18). On Day 16, a challenge dose of 2 x 10' cells ofthe K1735 cell lines or 2 x IO6cells of the SF19 cell lines was given

s.c. on the opposite side. These challenge doses produce tumors in100% of the mice challenged. Mice were checked weekly for tumorgrowth for 6-10 weeks.

In Vitro Th Cell Assay. To generate tumor-immune T-helper cells,C3H mice were immunized with three i.p. injections of 2 x IO6 X-irradiated (10,000-rad) tumor cells at 1-week intervals. Spleens fromthese mice were used as a source of tumor-immune Th cells 2 weeksafter the third immunization.

Concurrently, other groups of mice were immunized with TNP-conjugated SRBC to generate TNP-SRBC-primed B-cells. SRBC inAlsever's solution were obtained from the Science Park Veterinary

Division of the University of Texas M. D. Anderson Cancer Center.TNP-conjugated erythrocytes were freshly prepared for each experiment. Red cells were washed four times in PBS, pH 7.4, beforeconjugation with TNP. This was accomplished by mixing, 1-ml packedvolume of washed SRBC and 7 ml of 20 mM TNP in PBS, pH 8.O.After 30 min at room temperature, the cells were washed four timeswith PBS, pH 7.4. Glycylglycine (10 mg/ml) was added to the firstwash. To prime mice against TNP-SRBC, they were given two i.p.injections of 0.1 ml of 2% (v/v) TNP-SRBC, 2 weeks apart. These micewere used as a source of TNP-primed B-cells 2 weeks after the secondimmunization.

Spleens were removed from tumor- and TNP-SRBC-immune mice,and single-cell suspensions were prepared. The cells were washed, thenresuspended in RPMI 1640, and the number of viable cells was determined by trypan blue exclusion.

B-cells were purified from mice immunized against TNP-SRBC.Spleen cells were suspended at 4 x 107/ml in RPMI containing 2%FBS, and an equal volume of RPMI containing anti-Thy-1.2 (Becton-Dickinson, Oxnard, CA) was added. The final dilution of the Thy-1.2monoclonal antibody was 1/166. After 30 min on ice, the cells werewashed twice and resuspended in RPMI with 2% FBS and a 1/16dilution of rabbit H2 complement (Pel-Freez, Brown Deer, WI). Thecells were incubated for l h at 37"( ' and washed twice before the number

of viable spleen cells was counted as above.Spleen cell preparations from tumor-immunized mice were enriched

for T cells by incubation on NW columns by the method of Julius etal. (19). The NW-nonadherent cells were washed and used as Th in themixed lymphocyte-tumor cell cultures.

Cultures were set up in RPMI 1640 containing 10% FBS, 2 mM L-glutamine, 5 x 10~5 M 2-mercaptoethanol, 100 U/ml penicillin, 100

iig/ml streptomycin, 10% sodium pyruvate, 10 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid buffer, and 1 x vitamins and nonessential amino acids (GIBCO). To test for Th activity, we used the methodof Howie and McBride (20). Cultures consisting of IO7 T-depletedspleen cells from TNP-SRBC-primed mice, 5 x 10* TNP-conjugated,X-irradiated (10,000-rad) tumor cells, and Kf NW-nonadherent spleencells from tumor-immunized mice were grown in 3 ml of medium inFalcon six-well tissue culture plates (Becton- Dickinson, Oxnard, CA).To determine the background PFC response for each experiment, B-cells were incubated alone and with TNP tumor cells.

After a 5-day incubation, the cultures were harvested and washed inHBSS. Anti-TNP PFC were assayed by using the Mishell and Dutton(21) modification of the .lerne and Nordin (22) technique. Each samplewas assayed in the presence of 8% TNP-HRBC, guinea pig complement(Pel Freez), and rabbit anti-mouse IgG serum (Cappel, Cochranville,PA). Triplicate determinations were done on each culture. The numberof anti-TNP PFC were expressed as total (direct plus indirect) plaquesper culture. Cultures were also tested for PFC against SRBC and HRBCto demonstrate the specificity of the antibody response for TNP.

In Vitro CTL Assay. C3H mice were immunized with one i.p.injection of 2 x 10' X-irradiated (10,000-rad) tumor cells to generatetumor-immune CTL. 2 weeks later, spleens were removed from mice,and single-cell suspensions were prepared. The cells were washed, thenresuspended in RPMI 1640, and the number of viable cells was determined by trypan blue exclusion.

As a source of T-enriched CTL, spleen cell preparations from tumor-immunized mice were incubated on nylon wool columns using themethod of Julius et al, (19). The NW-nonadherent cells were washedand used as CTL in the mixed lymphocyte-tumor cell cultures.

Cultures were set up in the same media as described above for theTh cell assays. Five x 10' NW-purified splenocytes and 10*X-irradiated(10,000-rad) tumor cells were incubated in 2 ml of medium in Falcon24 «elltissue culture plates for five days (Becton-Dickinson, Oxnard,CA).

To assay for CTL activity, we adapted the method described byThorn (9). Tumor targets were labeled with Na2"CrO4 immediatelyafter trypsinization, and 100-^1 aliquots of CTL and "Cr-labeled tumortarget cells were added to each well of a V-bottom microtest plate(Falcon). The plates were centrifuged for 4-5 min at 400 RPM andincubated for 4 h at 37°C.Cytotoxicity was determined by sampling

100 n\ directly from each well. The percentage of cytotoxicity wascalculated as [(E-S)/(T-S)\ x 100%, where E is the counts released inthe presence of CTL cells, 5 is the counts released by targets incubatedonly with medium, and T is the counts released by target cells in thepresence of 1% Triton X-100. The total release (T) was at least 75% ofthe incorporated counts for all target cells. The spontaneous release (S)was consistently <15% of releasable counts.

Chronic UV Irradiation of Mice. Mice were exposed to UVR according to the procedure of Kripke (17). The light source was a bank of sixFS40 sunlamps (Westinghouse). Dorsal hair of the mice was removedwith electric clippers once a week, and mice were exposed to UVR for

1208

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIGENS ON UVR-INDUCED ANTIGENIC VARIANTS

1 h, three times a week, for at least 12 weeks. The incident dose ratewas approximately 4 W/m2 over the wavelength range from 280 to 320

nm. None of the animals had yet developed primary tumors from theUV irradiation at the time of these experiments.

RESULTS

In Vivo Cross-Reactivity between the Antigenic Variants andParental Cell Lines. An in vivo cross-protection experiment wasperformed to determine whether the immune responses generated by immunization with the antigenic variants were protective against challenge with the parent tumor. Nonimmunogenicclones of UV-irradiated parental cultures and other antigenicvariants were also tested for cross-reactivity. Tables l (SF19cell lines) and 2 (K1735 cell lines) show that mice immunizedwith one antigenic variant cell line generated an immune response that resulted in the rejection of the immunizing antigenicvariant as expected. In addition, the variant-immune mice were

able to reject a challenge with the nonimmunogenic parentalcells, and the nonimmunogenic clones, SF19-LD.3 (Table 1)and K1735-OR.3 (Table 2), derived from UV-irradiated SF19and K1735 cells, respectively. However, animals immune to theantigenic variant did not reject a challenge with a differentantigenic variant or an unrelated control tumor. Mice immunized with cells of the parent tumors or nonimmunogenic clonesgenerated no detectable resistance to tumor challenge. Control,X-irradiated, unimmunized mice were unable to reject a challenge with any of the cell lines.

The lack of cross-protection among antigenic variants indicates that the variants exhibit individually specific, non-cross-

Table 1 Cross-protection between SF19 fibrosarcoma and antigenic variant celllines

Immunizingcellline0SF19

SF19-OR.26SF19-OR.58SF19-LD.3No immunizationNo.

of mice with tumor/no. of miceinjected*SF1913/15

0/151/158/10

10/10SF19-OR.2613/15

1/1514/149/10

10/10SF19-OR.5815/15

15/150/15

10/1010/10SF19-LD.310/10

1/100/10

10/1010/10MCA11310/10

10/1010/1010/1010/10

"C3H mice were immunized by s.c. injection of 2 x 10* X-irradiated tumor

cells on Day 0. Control, unimmunized mice were injected with an equal volumeof culture medium. On Day 14, the mice were exposed to 450 rads of X-irradiation. On Day 16, a challenge dose of 2 x 10*cells of the appropriate tumor

cell line was given s.c. on the opposite side. Mice were checked weekly for tumorgrowth for 10 weeks.

"The antigenic variants, SF19-OR.26 and SF19.OR.58, and the nonimmunogenic clone, SF19-LD.3, are clones derived from UV-irradiated SF19 cultures.MCA113 is a methylcholanthrene-induced fibrosarcoma used as a negative control. Data are reported for 10 weeks after challenge.

Table 2 Cross-protection between Kl 735 melanoma and antigenic variant celllines

Immunizingcellline'K1735

K1735-OR.5K1735-LD.1K1735-OR.3No immunizationNo.

of mice with tumor/no. of miceinjected*K

173514/15

2/151/15

10/1010/10K1735-OR.512/15

0/1513/1510/1010/10K1735-LD.114/15

15/151/159/10

10/10K1735-OR.310/10

1/102/10

10/1010/10MCA11310/10

10/1010/1010/1010/10

°C3H mice were immunized by s.c. injection of 2 x 10* X-irradiated tumorcells on Day 0. Control unimmunized mice were injected with an equal volumeof culture medium. On Day 14, the mice were exposed to 450 rads of X-irradiation. On Day 16, a challenge dose of 2 x 10*cells of the appropriate tumor

cell line was given s.c. on the opposite side. Mice were checked weekly for tumorgrowth for 8 weeks.

* The antigenic variants, K1735-OR.5 and K1735-LD.1, and the nonantigenicclone, K1735-OR.3, are clones derived from UV-irradiated K1735 cultures.M CAI 13 is a methylcholanthrene-induced fibrosarcoma used as a negative control. Data are reported for 8 weeks after challenge.

reacting antigens. This individual specificity of the variant celllines resembles that of skin cancers induced in mice in vivo byrepeated exposure to UVR (2). The presence of cross-reactivitybetween an antigenic variant and parental cells and cells of anonimmunogenic clone indicates that the antigenic variants,parental cells, and nonimmunogenic clones share at least onetumor rejection antigen that by itself is not sufficient to inducea protective immune response.

Specificity of Th Cells and CTL from Tumor-immunized Mice.Immune responses to tumor antigens, similar to responsesagainst other antigens, involve a complex immune network thatincludes B-cells, various subsets of T-cells, and macrophages.This network includes antigen recognition (afferent) and anti-tumor effector (efferent) phases. Therefore, when evaluatingthe specificity of the immune rejection response to any tumor,it is important to define specificity in terms of the variousaffector and effector cell types. In this study, the specificity ofantigenic variant-induced Th cells (affector) and CTL (effector)was determined in vitro and compared with the specificity ofthe antitumor rejection response observed in vivo.

In order to determine whether immunization with the antigenic variants induces Th activity, an indirect in vitro assay wasperformed. In addition, the specificity of the Th cells wasassessed by determining their ability to recognize unique and/or shared antigens present on the parental cells and the otherclonal cell lines derived from the UV-irradiated parent cell line.

Four tumor lines (K1735, K1735-OR.5, K1735-LD.1, andK1735-OR.3) were injected into different groups of mice togenerate Th. C3H mice were injected i.p. three times weeklywith X-irradiated (10,000-rad) tumor cells. Splenic NW-non-adherent cells from the tumor-immunized mice were coculti-vated for 5 days in vitro with TNP-primed B-cells and TNP-conjugated tumor cells of the appropriate cell line. The resultsshown in Table 3 demonstrate that the amount of Th activityinduced by the antigenic variants was much higher than thatmeasured after immunization with the cells of the parentaltumor or nonimmunogenic clone. The antigenic variant-induced Th cells recognized the immunizing variant, the parenttumor, and the nonimmunogenic clone, but did not cross-reactwith a different antigenic variant or unrelated, control tumorcells. These results suggest that Th cells generated by immunization of mice with antigenic variant cells detect variant-specificantigens as well as antigens present on the parental and nonimmunogenic clone cells. This experiment was repeated andyielded the same results.

Splenic T cells from mice immunized with tumor cells weretested for reactivity against a battery of target cells in a "Cr-

release CTL assay. Table 4 shows that CTL from mice immunized with antigenic variant cells recognized the immunizingvariant cells, the parent cells, and the nonimmunogenic clonecells, but not the other antigenic variant cells. The amount ofcytotoxicity against the parent tumor and nonimmunogenicclone was less than that against the immunizing variant. Theseresults suggest that there are CTL-defined antigens on thevariant cells specific for the variant, as well as antigens sharedwith the parental and nonimmunogenic clone cells. Immunization with cells of the parent tumor or nonimmunogenic clonegenerated a small amount of CTL activity, but this was notgreater than that generated against an unrelated tumor. Becausethis activity is nonspecific, it may represent a low level of naturalkiller cell activity. This experiment was performed twice withthe same results.

Growth of an Antigenic Variant in UV-irradiated Mice. Todetermine whether antigenic variants derived from K1735 mel-

1209

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIGENS ON UVR-INDUCEDANTIGENIC VARIANTS

Table 3 Specificity of helper T-lymphocytes for individual Kl 735 tumor lines

T-helper cellsimmunizedto"K

1735K1735-OR.5K1735-LD.1K1735-OR.3No

immunizationTNP-conjugated

tumorusedTNP-K1735131

±IIe1419

±791329±137171

±721±9TNP-OR.5113±

151735±64177±40210

±13630±5TNP-LD.l169

±27131±132068

±84206±3727±4in

vitro11TNP-OR.3192

±121181±1711206±150148

±918±7TNP-MCA11319±

140±625±510±924

±8" Normal C3H mice were immunized i.p. three times at weekly intervals with 2 x 10' cells of the appropriate tumor. 1 week after the third immunization, NW-

nonadherent spleen cells were isolated from tumor-immunized and nonimmune mice and test for Th activity in vitro. The NW-purified Th cells were cocultivated invitro for 5 days with TNP-SRBC-primed B-cells and TNP-conjugated tumor cells.

* K1735 is the parental melanoma cell line. K1735-OR.5 and -LD.l are antigenic variants and K1735-OR.3 is a nonimmunogenic clone derived from I'VR-¡rradiati-i!K1735 cultures. MCA113 is a methylcholanthrene-induced fibrosarcoma used as a negative control.

e Values represent the number of plaque-forming cells (±SE) generated in cultures containing IO7tumor-primed Th cells, IO7TNP-primed B-cells, and 5 X 10sTNP-tumor cells.

CTL from miceimmunizedwith"K1735

K1735-OR.5K1735-LD.1K1735-OR.3No immunizationTable

4Specificity of CTL for individual Kl 735 tumorlinesCytotoxicityof tumor targets' used invitroK173512.5±4.1C

42.3 ±5.453.6 ±9.016.3 ±5.80.0 ±0.0K1735-OR.512.9

±6.270.7 ±9.915.0 ±5.213.7 ±3.92.0 ±3.2K1735-LD.117.3

±3.719.9 ±5.467.5 ±13.317.9 ±6.6

1.5 ±4.3K1735-OR.314.0

±2.536.4 ±3.141.6 ±18.213.9 ±5.50.8 ±0.9MCA11312.0

±4.912.8 ±4.210.1 ±5.613.0 ±6.14.2 ±5.6

°Nylon wool-purified spleen cells from mice immunized with 2x10' cells of the appropriate tumor cell line 2 weeks prior to in vitro culturing. A 4-h, "Cr-releaseassay was done as described in "Materials and Methods." Effector-to-target ratio was 100:1.

*K1735 is the parental melanoma cell line. K1735-OR.5 and -LD.l are antigenic variants and K1735-OR.3 is a nonimmunogenic clone derived from UVR-irradiated K1735 cultures. MCA 113 is a methylcholanthrene-induced fibrosarcoma used as a negative control.

' Values, percentage cytotoxicity ±SE.

anoma express the common, UVR-associated antigenic determinant recognized by UVR-induced Ts cells, K1735-LD.1 wasinjected into chronically UV-irradiated mice. As a positivecontrol, the UV-2240 fibrosarcoma was also injected into mice.This tumor cell line grows in UV-irradiated and immunosup-pressed mice but not in untreated mice. In addition, mice werechallenged with the MCA-FM1 fibrosarcoma, an antigenicvariant derived from a tumor induced in a C3H mouse byinjection of the chemical carcinogen 3-methylcholanthrene andtreated with the mutagen MNNG in vitro. MCA-FM1 cellsgrow in immunosuppressed but not in normal mice.The results,shown in Table 5, indicate that the K1735 antigenic variantLD.l, like the UV-2240 cells, grew in UV-irradiated and immunosuppressed mice, but not in untreated mice. These resultsare the same as those of previous experiments utilizing antigenicvariants derived from UV-irradiated SF19 fibrosarcoma cells(13). In contrast, MCA-FM1 cells grew only in the immunosuppressed mice but not in normal or UV-irradiated mice. Theparental K1735 cell line produced tumors in all groups of mice.Therefore, the K1735-LD.1 cells exhibited a pattern of growthunique to tumors bearing the UVR-associated antigen recog-

Table 5 Growth ofKl 735 melanoma-derived antigenic variant cells in normal,W-irradiated, and immunosuppressed mice

Tumor cellline"K1735

(parental)K1735-LD.1UV-2240MCA-FM1Normal

mice5/5

0/50/50/5No.

of mice with tumor/no. of miceinjectedUVR-irradiated

Immunosuppressedmice*mice'5/5

4/44/5 4/45/5 4/40/5 4/4P<NS"

0.0250.004NS

"C3H mice were injected with 2 X 10' K1735, K1735-LD.1, or MCA-FM1cells and 2 x 10' UV-2240 cells s.c., according to their minimal tumorigenicdose. The mice were checked weekly for tumor growth for 6-8 weeks.

* Exposed to su»lainp irradiation for 1 h, three times per week for 12 weeks.c Thymectomy at 4 weeks of age, followed by whole-body exposure to 450-rad

X-irradiation.'' Fisher's exact test utilized to compare growth of tumor cells in normal mice

versas UV-irradiated mice; NS, nonsignificant.

nized by UVR-induced Ts cells. The rejection of the MNNG-induced antigenic variant MCA-FM1 in UV-irradiated micesuggests that the expression of the UVR-associated commonantigen may be restricted to antigenic variants induced by UVR.

DISCUSSION

In these studies, we investigated whether the antigenic variants generated from UV-irradiated, nonimmunogenic murinetumor cellsexhibit the UVR-associated antigen unique to UVR-induced tumors and whether the variants share antigens withthe parental cells and other clones derived from UV-irradiatedcultures.

Our study demonstrates that in vitro UV irradiation of aweakly antigenic melanoma cell line generates antigenic clonesthat exhibit the UVR-associated common antigenic determinant recognized by the Ts cells induced in chronically UV-irradiated mice. In our experience, this antigen is found onlyon skin cancers induced in vivoor in vitro by UVR (4, 6, 23-25). Previously, we reported the presence of the UVR-associated antigen on antigenic variants derived from a spontaneousC3H strain fibrosarcoma (13). The induction of the UVR-associated antigen on a weakly antigenic tumor cell by in vitroexposure to UVR implies that at least some UVR-associatedantigens arise as a consequence of exposing cells to UVR andthat they may occur independently of the initial neoplastictransformation event.

The results of our in vivoand in vitro studies indicate that inaddition to the UVR-associated antigen, the antigenic variantsexhibit both unique, variant-specificantigens and parent-tumor-derived antigens. The antigenic variants produce a high level ofprotective immunity in vivo,which enables the murine host toreject the immunizing variant, the parental cells, and cells ofthe nonimmunogenic clone. These results indicate that theantigenic variants express a determinant that is also expressedby the parental and nonimmunogenic clone cells. This determinant seems to be nonimmunogenic by itself, but very immu-

1210

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIGENS ON UVR-INDUCED ANTIGENIC VARIANTS

nosensitive. Despite the apparent inability of the parental ornonimmunogenic cells to induce a detectable immune response, they are nevertheless rejected by an immune responseinduced by the antigenic variants. The specificities of Th andCTL in the in vitro assays are identical to the specificity of theimmune response observed in vivo, suggesting that these cellsmay be important in the immune-mediated resistance to tumorchallenge.

The ability of the variants to cross-react with the parentalcells is consistent with the results of other investigators usingdifferent mutagenic agents (26-33). In most of these studies,the variant lines induced a weak immune protection againstother antigenic variant lines (26-32), and Peppoloni et al. (33)reported that UVC radiation (254 nm)-induced antigenic variants exhibited complete interclonal cross-reactivity. In ourstudies in which UVB radiation was used, mice immunized withone variant were not able to reject a challenge with cells ofanother variant. Therefore, no evidence of cross-reactivity between antigenic variants in vivo was found (Tables 1 and 2).This lack of cross-reactivity is not attributable to our use of X-irradiated cells for immunization because identical results wereobtained in studies using non-X-irradiated cells (13). The reasonfor this difference between antigenic variants induced with UVCand UVB radiation is not clear, but it may be due to individualproperties of the cell lines used, to shared viral antigens, or tothe fact that various wavelengths of UVR have different biological effects. The individual specificity of variant cell lines produced by UVB radiation resembles that of skin cancers inducedin mice in vivo by repeated exposure to UVB radiation, as testedby both in vivo (1,3) and in vitro (8, 9, 11, 12, 34) assays.

The lack of detectable in vivo and in vitro cross-reactivitybetween the antigenic variants in this study is interesting inlight of the fact that in the same assays, the variants shareantigens with all other cell lines derived from the parent, i.e.,parent cells themselves and the nonimmunogenic clone cells.From these observations, it appears that the presence of aunique, variant-specific antigen on an antigenic variant permitsthe induction of a tumor rejection response against all tumorsderived from the parental cells unless a different, dominant,variant-specific antigen is exhibited by the cells. In support ofthis theory, Wortzel et al. have identified multiple antigens ona tumor induced in mice by UVR, which are expressed in ahierarchical manner (11, 12). Thus, at least for one UVR-induced tumor, the immune system responds to different combinations of antigens present on tumor cells in a specific andunique way.

The structural arrangement of the antigens induced in theparent tumor cells by exposure in vitro to UVR is not known.However, based on the results of these experiments, two modelscan be proposed. Evidence from studies in other systems havesuggested that the immune response against a weakly immu-nogenic determinant can be elicited only if another (helper)determinant is corecognized on the same cell (35,36). Likewise,UVR may induce a new determinant on native, nonimmunogenic, cell surface molecules thereby producing an "altered self"

molecule that is now recognizable by the immune system. Thesedeterminants may occur on the same (Fig. \A) or different (Fig.1B) cell surface molecules exhibited on a single cell. In the firstmodel (Fig. 1/1), exposure to UVR of a parental cell exhibitinga nonimmunogenic cell surface molecule results in a new,unique epitope(s) on this molecule. Each antigenic variantexpresses its own neoepitope, as well as the nonimmunogenic(parent) molecule, whereas the nonimmunogenic clones expressno new epitopes on the common molecule. In addition, the

Antigenic

. io <2>

Parental cell

UV-induMdcelli II

UV inducid UV-lndue«d

Fig. 1. Models for the arrangement of the antigenic determinants present onUVR-induced antigenic variants. According to these models, IV radiation induces new epitopes on native, nonimmunogenic cell surface molecules. The newepitopes induced by UV radiation may be exhibited on (A) the same or (IS)different molecules present on the surface of the tumor cells exposed to UVradiation. In addition, UV radiation induces the expression of the UVR-associatedcommon determinant on the cells.

UVR-associated common antigen that is recognized by UVR-induced suppressor cells is expressed by the UVR-exposedparent cells. Immunization with an antigenic variant results inimmune effector cells that recognize the neoepitope alone, theneoepitope/parent molecule in combination, and the parentalmolecule by itself. Cells that recognize the neoepitope mediaterejection of the immunizing antigenic variant and are thereforevariant-specific immune cells. The cells that recognize the neo-epitope/parental combination or the parent molecule by itselfmediate the cross-reactive immunity that results in the rejectionof the parental and nonimmunogenic clones. These same effector cells specific for the parental antigen do not recognize the

1211

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIGENS ON UVR-1NDUCED ANTIGENIC VARIANTS

parental molecule on other antigenic variants because the parent molecule has been altered by the neoepitope unique to otherantigenic variants. The presence of a different neoantigen maychange the conformation of the parental molecule or stericallyhinder its recognition such that parent antigen-specific effectorcells no longer recognize the parent molecule.

This model is based on the assumption that the mutageniceffect of UVR resulting in increased antigenicity selectivelyalters one particular gene encoding a specific cell surface protein. In support of a selective mutation by UVR, bacterialmutagenesis studies have shown that UVR produces lesions inonly a few specific places in bacterial DNA (37), and studies byBohr et al. (38) suggest that repair of UVR-induced lesions maybe selective for certain genes and not others. The major problemwith this model is, however, that UVR-induced skin cancers donot seem to exhibit their unique antigens on the same cellsurface molecule. The unique antigen on the UVR-inducedparent 1591 is an altered histocompatibility antigen (39),whereas a different UVR-induced tumor (UV-2240) possessesunique antigens that are biochemically extractable with 1-bu-tanol, a solvent that does not extract major histocompatibilitycomplex antigens (40). In addition, the expected frequency ofmultiple mutations in the same gene would be exceedingly low,resulting in a very low frequency of antigenic variants. However,a high frequency of antigenic variants is produced by in vitroUVR irradiation.

Alternatively, UVR may induce the expression of uniqueantigenic determinants on different cell surface molecules asshown in Fig. IB. In this model, following UV irradiation ofthe parental tumor cultures, the epitope unique for one antigenic variant is present on a cell surface molecule that isdifferent from the molecule bearing the unique epitope of asecond variant. UVR also induces the expression of the UVR-associated common antigen on the tumor cells. As in the firstmodel, immunization with one antigenic variant would produceeffector cells that recognize the unique epitope alone, the neoepitope in combination with a particular parent molecule, andthe parental molecule by itself. Effector cells generated byimmunization with a different antigenic variant would be specific for both a different unique epitope and a different parentmolecule. The main problem with this model is that if effectorcells can be generated that recognize the original parentalmolecules, why then do they not recognize and react againstthe same parent molecule present on another antigenic variantcell? One possibility is that the parent molecule exhibiting theneoantigen blocks the binding of the effector cells specific forother parent molecules by sterically hindering or allostericallychanging the conformation of these molecules on the cell surface. It may even be that the expression of the molecule bearingthe neoantigen somehow leads to a decrease in the expressionof other parent molecules on the cell surface. In light of thehigh frequency of antigenic variants produced in vitro by UVR,this second model may be more likely than the first, becauseexposure of a cell to UVR probably produces mutations in avariety of genes rather than multiple mutations in a single,exclusive gene. In this model, UVR may still produce lesionsin selective regions of the genome as described for the firstmodel, but no conclusive evidence exists that only one genewithin this selected subset of genes is targeted for mutation.Finally, it is not yet clear whether the UVR-associated antigenand the tumor-specific antigen are induced in the same or

separate cell surface molecules.These models of course do not take into account other

possible tumor-associated antigens, such as suppressogenic de

terminants and determinants recognized by other effector cells(macrophages, natural killer cells, and lymphokine-activatedkiller cells). All of these may be involved in the complex reactionthat results in selective recognition and rejection of the tumorcells. Additional analysis of the immunogenic determinantspresent on the antigenic variants at a molecular level, as wellas analysis of the specificity of cloned affector/effector cells, isnecessary to distinguish between the two models proposedabove.

Our studies demonstrate that in vitro exposure of nonanti-genic tumors to UVB radiation induces antigenic variants thatexhibit antigenic characteristics similar to that expressed byUVR-induced tumors and are capable of immunizing against achallenge with the parental tumor from which the variants werederived. The ability to produce in vitro a highly antigenic variantthat is recognized and rejected by the immune system and thatalso cross-reacts with the parental nonantigenic tumor is ofconsiderable importance for the immunotherapy of cancer. Asa result of the antigenic cross-reactivity between UVR-inducedtumor variants and the parent tumor cells, tumors that elicit apoor immune response because of low antigenicity could beremoved surgically, established in short term culture, exposedto UVR in vitro, and used for immunization against recurrenceof the primary tumor or métastases.

ACKNOWLEDGMENTS

The secretarial assistance of Alice Burnett and the technical assistance of Jacquelyn Walker is acknowledged with thanks.

REFERENCES

1. Kripke, M. L. Antigenicity of murine skin tumors induced by ultravioletlight. J. Nati. Cancer Inst., 53: 1333-1336, 1974.

2. Kripke, M. L., and Fisher, M. S. Immunologie parameters of ultravioletcarcinogenesis. J. Nati. Cancer Inst., 57: 211-215, 1976.

3. Pasternak, G., Graffi, A., and Horn, K. H. Der Nachweis individual spezifischen Antigenitätder UV-induzierten Sarkomen der Maus. Acta Biol. Med.Ger., 13: 276-279, 1964.

4. Fisher, M. S., and Kripke, M. L. Systemic alteration induced in mice byultraviolet light irradiation and its relationship to ultraviolet carcinogenesis.Proc. Nati. Acad. Sci. USA, 74: 1688-1692, 1977.

5. Spellman, C. W., and Daynes, R. A. Modification of immunologie potentialby ultraviolet radiation. II. Generation of suppressor cells in short-term UV-irradiated mice. Transplantation, 24: 120-126, 1977.

6. Kripke, M. L., Thorn, R. M., Lili, P. H., Civin, C. I., Fisher, M. S., andPazmino, N. H. Further characterization of immunologie unresponsivenessinduced in mice by UV radiation: growth and induction of non-1 V inducedtumors in UV-irradiated mice. Transplantation, 28: 212-217, 1979.

7. Roberts, L. K. Characterization of a cloned ultraviolet radiation (UV>induced suppressor T cell line that is capable of inhibiting anti-UV tumorimmune responses. J. Immunol., 136: I908-19I6, 1986.

8. Fortner, G. W., and Kripke, M. L. In vitro reactivity of splenic lymphocytesfrom normal and UV-irradiated mice against syngeneic UV-induced tumors.J. Immunol., 118:1483-1487, 1977.

9. Thorn, R. T. Specific inhibition of cytotoxic memory cells produced againstUV-induced tumors in UV-irradiated mice. J. Immunol., 121: 1920-1926,1978.

10. Romerdahl, C. A., and Kripke, M. L. Regulation of the immune responseagainst UV-induced skin cancers: specificity of helper cells and their susceptibility to UV-induced suppressor cells. J. Immunol., 137: 3031-3035,1986.

11. Wortzel, R. D., Phillips, C., and Schreiber, H. Multiple tumor-specificantigens expressed on a single tumor cell. Nature (Lond.). 304: 165-167,1983.

12. Wortzel, R. D., Urban, J. L., and Schreiber, H. Malignant growth in thenormal host after variant selection in vitro with cytolytic T-cell lines. Proc.Nati. Acad. Sci. USA, 81:2186-2190, 1984.

13. Hosteller, L. W., Ananthaswamy, H. N., and Kripke, M. L. Generation oftumor-specific transplantation antigens by UV radiation can occur independently of neoplastic transformation. J. Immunol., 137: 2721-2725, 1986.

14. Hosteller, L. W., and Kripke, M. L. Proportion of antigenic variants inducedby in vitro UV irradiation differs in clones derived from a single tumor. J.Immunol., 140: 666-670, 1988.

15. Kripke, M. L. Speculations on the role of ultraviolet radiation in the development of malignant melanoma. J. Nati. Cancer Inst., 63: 541-548, 1979.

16. Fidler, I. J., Gersten, D. M., and Kripke, M. L. Influence of immune status

1212

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

ANTIGENS ON UVR-INDUCED ANTIGENIC VARIANTS

on the metastasis of three murine fibrosarcomas of different immunogenici-ties. Cancer Res., 39: 3816-3821, 1979.

17. Kripke, M. !.. Latency, histology and antigenicity of tumors induced byultraviolet light in three inbred mouse strains. Cancer Res., 37: 1395-1400,1977.

18. Brent, I... and Medawar, P. B. Quantitative studies on tissue transplantationimmunity. 8. The effects of irradiation. Proc. R. Soc. Lond. (Biol.), 165:413-423, 1966.

19. Julius, M. II., Simpson, E., and Herzenberg, L. A. A rapid method for theisolation of functional thymus-derived murine lymphocytes. Eur. J. Inimii-nol., 3: 645-649, 1973.

20. Howie, S., and McBride, W. H. Tumor-specific T helper activity can beabrogated by two distinct suppressor cell mechanisms. Eur. J. Immunol., 12:671-675, 1982.

21. Mishell, R., and Dutton, R. W. Immunization of dissociated spleen cellcultures from normal mice. J. Exp. Med., 126:423-442, 1967.

22. .lerne. N., and Nordin, A. Plaque formation in agar by single antibody-producing cells. Science (Wash. DC), 140:405, 1963.

23. Fisher, M. S., Kripke, M. L., and Chan, G. L. Antigenic similarity betweencells transformed by ultraviolet radiation in vitro and in vivo. Science (Wash.DC), 223: 593-594, 1984.

24. Fisher, M. S., and Kripke, M. L. Further studies on the tumor-specificsuppressor cells induced by ultraviolet radiation. J. Immunol., 121: 1139-1144, 1978.

25. Roberts, L. K., Lynch, D. H., and Daynes, R. A. Evidence for two functionallydistinct cross-reactive tumor antigens associated with ultraviolet light andchemically induced tumors. Transplantation, 33: 352-360, 1982.

26. Boon, T., and Van Pel, A. Teratocarcinoma cell variants rejected by syngeneicmice: protection of mice immunized with these variants against other variantsand against the original malignant cell lines. Proc. Nati. Acad. Sci. USA, 75:1519-1523, 1978.

27. Van Pel, A., Georlette, M., and Boon, T. Tumor cell variants obtained bymutagenesis of a Lewis lung carcinoma cell line: immune rejection bysyngeneic mice. Proc. Nati. Acad. Sci. USA, 76: 5282-5285, 1979.

28. Uyttenhove, <'.. Van Snick, J., and Boon, T. Immunogenic variants obtained

by mutagenesis of mouse mastocytoma P815.1. Rejection by syngeneic mice.

J. Exp. Med., 152: 1175-1183, 1980.29. Van Pel, A., and Boon, T. Protection against a nonimmunogenic mouse

leukemia by an immunogenic variant obtained by mutagenesis. Proc. Nati.Acad. Sci. USA, 79:4718-4722, 1982.

30. Van Pel, A., Vessiere, F., and Boon, T. Protection against two spontaneousmouse leukemias conferred by immunogenic variants obtained by mutagenesis. J. Exp. med., 157: 1992-2001, 1983.

31. Boon, T. Antigenic tumor variants obtained with mutagenesis. Adv. CancerRes., 39: 121-151, 1983.

32. Frost, P., Kerbel, R. S., Bauer, E., Tartamella-Biondo, R., and Cefalu, W.Mutagen treatment as a means for selecting immunogenic variants fromotherwise poorly immunogenic malignant murine tumors. Cancer Res., 43:125-132, 1983.

33. Peppoloni, S., Herberman, R. B., and Gorelik, E. Induction of highly immunogenic variants of Lewis lung carcinoma tumor by ultraviolet irradiation.Cancer Res., 45: 2560-2566, 1985.

34. Romerdahl, C. A., and Kripke, M. L. Role of helper T-lymphocytes inrejection of UV-induced murine skin cancers. Cancer Res., 48: 2325-2328,1988.

35. Keene, J., and Forman, J. Helper activity is required for the in vivogenerationof cytotoxic T lymphocytes. J. Exp. Med., 155: 768-782, 1982.

36. Ciavarra, R. P., Vitetta, E. S., and Forman, J. Growth inhibition of a B cellleukemia: evidence implicating an anti-idiotype immune response for protective tumor immunity. J. Immunol., 137: 1371-1375, 1986.

37. Miller, J. H. Mutagenic specificity of ultraviolet light. J. Mol. Biol., 182:45-65, 1985.

38. Bohr, V. A., Okumoto, D. S., and Hanawalt, P. C. Survival of UV-irradiatedmammalian cells correlates with efficient DNA repair in an essential gene.Proc. Nati. Acad. Sci. USA, 83: 3830-3833, 1986.

39. Phillips, C., McMillan, M., Flood, P., Murphy, D. B., Forman, J., Lancki,D., Womack, J. E., Goodenow, R. S., and Schreiber, H. S. Identification ofa unique tumor-specific antigen as a novel class I major histocompatibilitymolecule. Proc. Nati. Acad. Sci. USA, «2:5140-5144, 1985.

40. Simcik, W. J., Kripke, M. L., Sheu, T-L., and LeGrue, S. J. Differentialextraction of tumor-specific antigens from two ultraviolet light-induced murine fibrosarcomas with the use of 1-butano!. Cancer Res., 48: 7004-7008,1988.

1213

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1989;49:1207-1213. Cancer Res   Lisa W. Hostetler, Cynthia A. Romerdahl and Margaret L. Kripke 

in Vivo and in VitroTumor Cell Variants Measured Specificity of Antigens on UV Radiation-induced Antigenic

  Updated version

  http://cancerres.aacrjournals.org/content/49/5/1207

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/49/5/1207To request permission to re-use all or part of this article, use this link

on May 3, 2020. © 1989 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from