Archives of Virology
Archives of Virology 62, 209--219 (1979)
© by Springer-Ver[ag 1979
Production of Interferon by Human Tumor Cell Lines By PATRICIA JAMESOI~ and S. E. GROSSBERO Department of Microbiology, The Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A. With 1 Figure Accepted July 2, 1979
Summary Fourteen continuous h u m a n cell lines, including nine derived from tumors and five from non-neoplastic tissues, produced interferon in response to induction with bluegongue virus (BTV), Newcastle disease virus (NDV), and poly (I) - poly (C) complexed with DEAE-dextran. The seven best interferon-producing ceil lines (one from a melanoma, five derived from carcinomas, and one SV40-virus-transformed kidney cell line) responded to at least one of the viral inducers with yields of interferon over 1000 units/m]. Because the HT-1376 bladder carcinoma cell line produced high yields of interferon in this survey, and is easily propagated, the optimal conditions for interferon production were investigated, using BTV as the inducer. Interferon yields in 59 inductions over a period of about two years consistently fell within a 6-fold range, and had a geometric mean tiger of about 2700 reference units (RU)/ml, representing the production of about 3 RU/10s cells. This yield is comparable to mean tigers of 1 to 10 RU/10 s cells obtained by others with h u m a n leukocytes, foreskin cell strains, or the l~amalva lymphoblastoid celI line. UV-inaetivated BTV at a multiplicity corresponding to 10 PFU/cell was as effective an inducer in the ttT-1376 cell line as the fully infectious virus at a multiplicity of 1 PFU/cell. The interferon produced by the HT-1376 epithelial cell line has characteristics similar to the interferon induced by poly (I) • poly (C) in human diploid fibroblasts. These studies clearly demonstrate t h a t m a n y different types of tumor-derived cells have the capacity to produce interferon, and t h a t some equal or surpass the efficiency of diploid cells.
Introduction Interferons have great potential value in treatment of severe viral infections and cancers because of three salient characteristics : i) their inhibition of replication of m a n y viruses (5, 15) ; if) their suppression of cell division (5, 13, 15) ; and iii) their enhancement of cytotoxic activity of lymphocytes (8, 15). Antitumor effects have
0304-8608/79/0062/0209/$ 02.20
210
PAT~ICIA J~wrESO~Zand S. E. G~OSSBE~G:
been observed after intramuscular interferon injection, even in h u m a n malignancies with a "large tumor volume" (t4, 39). The evaluation of this antiviral and aatitumor therapeutic potential has been greatly restricted b y the lack of sufficient quantities of adequately purified interferon for clinical trials (5). An alternative approach for the use of interferon as an antitumor agent is to induce interferon production in the malignant cells themselves in situ, since endogenous levels of interferon are usually higher than any obtained by injection of interferon. However, most tumor cell lines in common laboratory use have been described as being generally poor producers of interferon: i) MOEH~I~G et al. obtained about 0.00037 units/cell from t u m o r cell fines or about 10 percent as much as from diploid cells in the same system (29), if) WO~THI~G~O~ and AA~O~SON (43) found t h a t cell cultures from five human tumors produced interferon titers of 3 to 100 nnits/ml in comparison with seven normal cell strains which yielded interferon titers of 100 to 1000 units/ml, and iii) BILLIAV et al. tested several cell lines derived from human tumors and obtained m a x i m u m yields of 500 reference units (RU)/ml with New. castle disease virus as the inducer and 3000 l~U/ml with superinduction techniques using p o l y ( I ) , poly(C), except for one cell line, ~G-63, which produced 30,000 R U / m l (3). The explanation for the poor response generally observed i n vitro might be either that malignant cells intrinsically produce little interferon or t h a t cell properties change from those of the original neoplastic tissue during prolonged cultivation in vitro (29, 33, 40, 43). Several lymphoblastoid cell lines, such as those derived from Burkitt's lymphoma patients, produce interferon of high titer (11,000 R U / m l from 107 cells/ml) (40) but these cells all contain multiple copies of at least part of the genome of Epstein-Barr virus (EBV) integrated into the cellular genome, even though they m a y not produce EBV either chronically or after treatment with inducers (35). Despite the problem E B V contamination poses, the production of interferon in clinically useful quantities b y these cell lines is being actively pursued (I1, 44). Others have developed a system for interferon production using an SV40-transformed human fibroblastoid cell line (2). Interferon produced b y cell lines derived from fibroblastie or epithelial cells m a y be espcially useful, since EI~O~i'~ and ST~AXDER (9) have demonstrated some tissue specificity in the action of interferon, as measured by the suppression of cell division. We tested a number of cell lines derived rather recently from tumors or nonneoplastic tissue i) to evaluate the interferon-producing capacity of transformed cells in vitro, and if) to identify potential continuous cell lines for the efficient, economical production of large amounts of well characterized interferon for possible use in antitumor or antivh-al therapy. Materials and Methods Cell Cultures
The following cell cultures were acquired and propagated as described previously (37) : the BUD-8 human adult skin fibroblast cell strain, the FL cell line from normal human amnion, the mouse L cell line and the rabbit 1~K-I3 ~ d n e y cell line. A human embryo brain g]ial cell line (HEG) was purchased from Flow Laboratories (Roekville, MD). The CaSki (34) and Dot cervical carcinoma and the NEWTovarian carcinoma eeN lines were donated by Dr. R. Patfllo of the Gynecology Department, The Medical
H u m a n Tumor Cell Lines Produce Interferon
211
College of ~qsconsin. A normal feline embryo cell line, F E A (25), was provided by Dr. O. Jarrett, University of Glasgow, Scotland. Dr. M. Ahmed, Pfizer Research L~boratories, Maywood, N J, provided the NC-37 (26), H E K (t), and KC (28) cell lines. Dr. S. A. Aaronson, l~ational Institutes of Helath (NIX), gave us the A 375 melanoma and A549 lung carcinoma cell lines (12). Dr. S. Rasheed, University of Southern California, Los Angeles, made available the HT-1417 lymphoblastoid cell line and the HT-1376 bladder carcinoma cell line (36). :NB cells (38), obtained from Dr. M. Oxman, University of California, San Diego, and H u F / A 2 8 cells (4), from Dr. J. Butel, Baylor College of Medicine, Houston, Texas, are SV40-transformed cell lines. The lymphoblastoid cell line 8392B (2t) was received from Dr. M. Essex, H a r v a r d University School of Public Health, Boston, MA. The MG-63 cell line (3) was provided by Dr. A. Billiau, University of Leuven, Belgium. H E K , KC, and H u F / A 28 cell lines were grown in minimal essential medium (MEM) with Earle's salts (EMEM), supplemented with 10 percent fetal bovine serum (FBS); the MEG, BUD-8, F E A and MG-63 cell cultures were all grown in the same medium except for addition of the noa-essential amino acids. The RTB, A549, and A375 cells were propagated in Dulbecco's modified MEM with glucose increased to 4.5 mg/ml and containing 10 percent FBS. W a y m o u t h MB 752/1 medium mixed with an equal volume of Gey's salt solution containing FBS was used to grow MEW, CaSki and Dot cells (10 percent FBS for H E W and 20 percent for the other lines). RPMI-1640 medium was used for the growth of other cells, supplemented with 10 percent FBS for 1N!C-37 and HT-1376 cells and 20 percent FBS for lgT-1417 and 8392B cells.
Viruses Bluetongue virus (BTV) serotype 10 (20), purchased from Colorado Serum Co. (Denver, CO) as a vaccine strain designated BT-8, was propagated in BHK-21 cells as described previously (24). Inactivation by ultraviolet (UV) irradiation was described previously (24). Sendal virus (NIX V 321-001-000) was propagated in chicken embryos and titrated in primary hamster kidney cell monolayers (16). Newcastle disease virus (NDV), CG strain, was prepared and titrated as described previously (30).
Inter]eron Induction Induction of interferon in lymphoblastoid cell lines followed the standard procedure described by CANTELL (6) for leukocyte cultures except t h a t we used RPML1640 with 5 percent serum. Other cells were induced by attaching virus to washed confluent monolayers of cells grown for approximately 4 to 6 days in 15× 125 m m screw-cap culture tubes, 30 ml plastic flasks, or 32 oz prescription bottles. At the end of the attachment period, a medium appropriate for the cell was added; the concentration of FBS was reduced to 2 percent, and the volmne was decreased to half that used for cell growth. For cells propagated in RPMI-1640 medium, McCoy's 5a or MEM supplemented with 2 percent FBS was used for interferon production. Supernatant fluids were collected 24 hours after the virus inoculation and stored at --70 ° C. Before titration of interferon, residual inducer virus was acid-inactivated as described previously (23).Polyriboinosinate. p o l y r i b o c y t i d y l a t e - - p o l y ( I ) , p o l y ( C ) - - ( P - L Biochemicals, Milwaukee, WI) was used to induce interferon in the manner described earlier (23).
Inter]etch Assay Interferon was quantitated by a hemagglutinin (HA) yield-reduction assay using encephalomyocarditis (EMC) virus (22) with the BUD-8 cell strain (37). Over the five-year period in which the screening of cell lines was conducted, the sensitivity of the interferon assay increased approximately three-fold. During the initial three years (n = 19) the assay measured about seven times the titer assigned to the I~TIH reference leukocyte interferon (G 023-901-527) ; in 1977 (n = 5) the assay measured about 20 times the titer assigned to the standard. The mean ratio for all titrations (n = 24, GMT i05-26 units/ml, SD = 10 °.3) of the standard suggests that, overall, this assay detects roughly 9 times its assigned value. The EMC HA yield-reduction assay measured about 15 Arch. Virol. 621~
212
PATRIClA J~.~ESO~ and S. E. G~oss]3E~G:
seven times the assigned potency of the N I H reference fibroblast interferon (G023-902527); a GMT of 68,000 units/ml was obtained in 15 titrations (SD = 1.7-fold). The relative sensitivity esti~nated from the ratio of observed/assigned titers of the fibroblast interferon standard was used for the adjustment of mean titers calculated from large numbers of samples, known to contain predominantly fibroblast-like interferon, to express their activities in terms of reference units (RU). A leukocyte interferon preparation used as a laboratory reference was included in every titration to detect assay v~ia bility. In addition to the observed difference in sensitivity of the assay, there is also a difference in the slopes of the dose response curves of leukocyte and fibroblast interferons as previously reported by ED¥ et al. (7), and confirmed in our own assay; in five titrations of the two NIH reference reagents the difference between slopes was small but highly significant (P ~ i0-3). Because of these differences in biological activities of the two types of interferons, and the fact that for most of the cells we tested, no determination was made as to the type of interferon produced, no adjustment of observed titers was made. Further, recent reports (18) indicate that any given cell type may produce leukocyte- and/or fibroblast-like interferons in varying proportions peculiar to the given cell.
Results
Inter/eron-Producing Ability o/ Diverse Cell Lines Because cells m a y v a r y in their capacity to produce interferon in relation to the inducer used, the interferon-producing ability of several cell lines was tested with one or more of three different inducers (Table 1): p o l y ( I ) , poly(C) complexed with DEAE-dextran, NDV, and BTV. The cell lines were derived from tumors or from presumably normal tissues intentionally infected with a transforming virus in vitro. All but two of the cell lines produced at least small amounts of interferon in response to at least one of the viral inducers. KC cells responded to p o l y ( I ) , poly(C) but not to NDV. The response noted for HuF/A28 is not large enough to distinguish it from the antiviral activity of diluted poly (I) • poly (C) t h a t might be carried over from the induced cell culture (see Table 1). Several lines responded well to only one of the inducers. The NC-37 line provided the best response of the three 15~nphoblastoid cells we tested. Another seven cell lines (iN:B, A549, A375, CaSki, Dot, H E W , and HT-1376) produced good yields of interferon ( ~ 1000 units/ml) in response to one or both of the viral inducers; four of these, three from carcinomas, produced more t h a n 6000 units/ml with a viral inducer. Among the lines which produced good yields of interferon, the IIT-1376 cell was the most easily propagated. Two cell lines derived from carcinomas (Dot and H E W ) produced especially high yields of interferon in response to BTV. However, these tumor cell lines grow very slowly and are difficult to propagate so t h a t their interferon-producing potential was not. studied further. Additional experiments on BTV induction of interferon in A375 cells resulted in interferon yields similar to those obtained in the initial screening tests (2000 to 5200 units/ml). The interferon production b y A375 cells remained constant with BTV multiplicities of 1.0, 10.0, or 100 PFU/cell, with which interferon yields of 5200, 3700, and 3700 units/mI, respectively, were obtained. However, UV-inactivated BTV (BTV-UV) at a multiplicity of 50 particles/cell induced only small amounts of interferon in A375 cells with a m a x i m u m of 900 units/ml.
H u m a n Tumor Cell Lines Produce Interferon
213
Table 1. InterJeron production by human cell linesa and strains Yield of interferon (units/ml) with the indicated inducer c Induced cell b
BTV
NDV
poly (I) • poly (C)
Normal cells BUD-8 skin strain t t E G embryo glial strain
50 nd
720 370
HeLa eeU derivatives F L amnion H E K embryo kidney
nd nd
68 75
nd 680
Virus in/ected cells *KC neuroblastoma astrocytes
nd
< 10
440
with integrated I~SV genome HuF/A 28 SV 40 transformed skin
nd a nd
~ I0
~ I0
15
(Faneoni anemia) NB SV40 transformed kidney
2,500
2,500
15
Lymphoblastoid cells *HT-1417 lymphoblastoid "8392B Iymphoblastoid NC-37 lymphoblastoid
25 200 1,200
25 15 1,200
170 210 970
Tumor derived cells *A549 lung carcinoma *A375 melanoma *CaSki cervical carcinoma *Dot cervical carcinoma
480 6,600 2,t00 7,000
2,500 200 2,100 3,000
15 15 nd nd
15,000 4,000
3,000 9,500
nd 1,600
*HEW ovarian carcinoma *HT-1376 bladder carcinoma
Cell lines derived from human tumors are designated by an asterisk b Each milliliter of interferon is the product of about 106 cells c Inducer virus doses: 20 to 50 PFU/cell. Poly (I) • poly (C), complexed with an equal weight or 10-fold excess of DEAE-dextran, was applied to cells at a concentration of i00 ~g/ml a Not done. Interferon induction by poly (I) • poly (C) was not tested in BUD-8 cells as done for other cells. To measure the antiviral effect which might be observed by simple transfer of the poly (I) • poly (C) used as an inducer in various cell cultures, BUD-8 cells were exposed for 24 hours to poly (I) • poly (C) dilutions in medium from I0 i~g/ml through 0.1 ~g/ml. An apparent antiviral titer of 2 was estimated from this control test to be possibly due to residual inducer.
Inter/eron Production by H T - 1 3 7 6 Cells A m o n g the cells we screened, the I:IT-1376 bladder carcinoma celt line appeared to be the most promising cell line for the p r o d u c t i o n of h u m a n interferon i n large q u a n t i t y . The cultures m a y be s u b d i v i d e d weekly a t a ratio of t : 4, reaching a cell d e n s i t y of a b o u t 2 × 104/cm u. I n t e r f e r o n i n d u c t i o n using conditions defined previously (23) resulted i n a geometric m e a n t i t e r (GMT) of 2600 R U / m l (GMT ---15"
214
PATI~ICIA JANESON
and S. E. G~OSSBERG:
18,500 units/ml in a 6-fold range of 10,000 to 62,000 units/ml) from 22 preparations obtained with cultures of about 2 × 107 cells in 32 oz bottles induced with B T V at a multiplicity of about 10 PFU/cell. Subsequently, a series of 37 inductions with a multiplicity of about 1 PFU/cell yielded a GMT of 2800 g U / m l (range of 6300 to 38,000 units/mt). I n the cumulative results of 59 inductions, the interferon yield from HT-1376 celt cultures approached 2700 t~U/ml or a p p r o x i m a t e l y 3 RU/103 cells. HT-1376 cells produced as m u c h interferon when induced with B T V - U V as with infectious B T V (Fig. 1); a multiplicity of 10 PFU/eell was used for the infectious virus, and the same particle dose was used for the inactivated BTV. Virus preparations irradiated until no infectivity was detectable still caused some cytopathie effect (CPE) ; further irradiation reduced the C P E to a negligible level without impairing the interferon-inducing capacity.
I00 z5 L) 50 ,~..4, 25
C
II!i,
Io 5
lo 4
~_ lOs
_--
8
E
\
o~ ~,.
103
106
"I-
.c_ •~
102 ;
104`
>
01
102
"C c
Archives of Virology 62, 209--219 (1979)
© by Springer-Ver[ag 1979
Production of Interferon by Human Tumor Cell Lines By PATRICIA JAMESOI~ and S. E. GROSSBERO Department of Microbiology, The Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A. With 1 Figure Accepted July 2, 1979
Summary Fourteen continuous h u m a n cell lines, including nine derived from tumors and five from non-neoplastic tissues, produced interferon in response to induction with bluegongue virus (BTV), Newcastle disease virus (NDV), and poly (I) - poly (C) complexed with DEAE-dextran. The seven best interferon-producing ceil lines (one from a melanoma, five derived from carcinomas, and one SV40-virus-transformed kidney cell line) responded to at least one of the viral inducers with yields of interferon over 1000 units/m]. Because the HT-1376 bladder carcinoma cell line produced high yields of interferon in this survey, and is easily propagated, the optimal conditions for interferon production were investigated, using BTV as the inducer. Interferon yields in 59 inductions over a period of about two years consistently fell within a 6-fold range, and had a geometric mean tiger of about 2700 reference units (RU)/ml, representing the production of about 3 RU/10s cells. This yield is comparable to mean tigers of 1 to 10 RU/10 s cells obtained by others with h u m a n leukocytes, foreskin cell strains, or the l~amalva lymphoblastoid celI line. UV-inaetivated BTV at a multiplicity corresponding to 10 PFU/cell was as effective an inducer in the ttT-1376 cell line as the fully infectious virus at a multiplicity of 1 PFU/cell. The interferon produced by the HT-1376 epithelial cell line has characteristics similar to the interferon induced by poly (I) • poly (C) in human diploid fibroblasts. These studies clearly demonstrate t h a t m a n y different types of tumor-derived cells have the capacity to produce interferon, and t h a t some equal or surpass the efficiency of diploid cells.
Introduction Interferons have great potential value in treatment of severe viral infections and cancers because of three salient characteristics : i) their inhibition of replication of m a n y viruses (5, 15) ; if) their suppression of cell division (5, 13, 15) ; and iii) their enhancement of cytotoxic activity of lymphocytes (8, 15). Antitumor effects have
0304-8608/79/0062/0209/$ 02.20
210
PAT~ICIA J~wrESO~Zand S. E. G~OSSBE~G:
been observed after intramuscular interferon injection, even in h u m a n malignancies with a "large tumor volume" (t4, 39). The evaluation of this antiviral and aatitumor therapeutic potential has been greatly restricted b y the lack of sufficient quantities of adequately purified interferon for clinical trials (5). An alternative approach for the use of interferon as an antitumor agent is to induce interferon production in the malignant cells themselves in situ, since endogenous levels of interferon are usually higher than any obtained by injection of interferon. However, most tumor cell lines in common laboratory use have been described as being generally poor producers of interferon: i) MOEH~I~G et al. obtained about 0.00037 units/cell from t u m o r cell fines or about 10 percent as much as from diploid cells in the same system (29), if) WO~THI~G~O~ and AA~O~SON (43) found t h a t cell cultures from five human tumors produced interferon titers of 3 to 100 nnits/ml in comparison with seven normal cell strains which yielded interferon titers of 100 to 1000 units/ml, and iii) BILLIAV et al. tested several cell lines derived from human tumors and obtained m a x i m u m yields of 500 reference units (RU)/ml with New. castle disease virus as the inducer and 3000 l~U/ml with superinduction techniques using p o l y ( I ) , poly(C), except for one cell line, ~G-63, which produced 30,000 R U / m l (3). The explanation for the poor response generally observed i n vitro might be either that malignant cells intrinsically produce little interferon or t h a t cell properties change from those of the original neoplastic tissue during prolonged cultivation in vitro (29, 33, 40, 43). Several lymphoblastoid cell lines, such as those derived from Burkitt's lymphoma patients, produce interferon of high titer (11,000 R U / m l from 107 cells/ml) (40) but these cells all contain multiple copies of at least part of the genome of Epstein-Barr virus (EBV) integrated into the cellular genome, even though they m a y not produce EBV either chronically or after treatment with inducers (35). Despite the problem E B V contamination poses, the production of interferon in clinically useful quantities b y these cell lines is being actively pursued (I1, 44). Others have developed a system for interferon production using an SV40-transformed human fibroblastoid cell line (2). Interferon produced b y cell lines derived from fibroblastie or epithelial cells m a y be espcially useful, since EI~O~i'~ and ST~AXDER (9) have demonstrated some tissue specificity in the action of interferon, as measured by the suppression of cell division. We tested a number of cell lines derived rather recently from tumors or nonneoplastic tissue i) to evaluate the interferon-producing capacity of transformed cells in vitro, and if) to identify potential continuous cell lines for the efficient, economical production of large amounts of well characterized interferon for possible use in antitumor or antivh-al therapy. Materials and Methods Cell Cultures
The following cell cultures were acquired and propagated as described previously (37) : the BUD-8 human adult skin fibroblast cell strain, the FL cell line from normal human amnion, the mouse L cell line and the rabbit 1~K-I3 ~ d n e y cell line. A human embryo brain g]ial cell line (HEG) was purchased from Flow Laboratories (Roekville, MD). The CaSki (34) and Dot cervical carcinoma and the NEWTovarian carcinoma eeN lines were donated by Dr. R. Patfllo of the Gynecology Department, The Medical
H u m a n Tumor Cell Lines Produce Interferon
211
College of ~qsconsin. A normal feline embryo cell line, F E A (25), was provided by Dr. O. Jarrett, University of Glasgow, Scotland. Dr. M. Ahmed, Pfizer Research L~boratories, Maywood, N J, provided the NC-37 (26), H E K (t), and KC (28) cell lines. Dr. S. A. Aaronson, l~ational Institutes of Helath (NIX), gave us the A 375 melanoma and A549 lung carcinoma cell lines (12). Dr. S. Rasheed, University of Southern California, Los Angeles, made available the HT-1417 lymphoblastoid cell line and the HT-1376 bladder carcinoma cell line (36). :NB cells (38), obtained from Dr. M. Oxman, University of California, San Diego, and H u F / A 2 8 cells (4), from Dr. J. Butel, Baylor College of Medicine, Houston, Texas, are SV40-transformed cell lines. The lymphoblastoid cell line 8392B (2t) was received from Dr. M. Essex, H a r v a r d University School of Public Health, Boston, MA. The MG-63 cell line (3) was provided by Dr. A. Billiau, University of Leuven, Belgium. H E K , KC, and H u F / A 28 cell lines were grown in minimal essential medium (MEM) with Earle's salts (EMEM), supplemented with 10 percent fetal bovine serum (FBS); the MEG, BUD-8, F E A and MG-63 cell cultures were all grown in the same medium except for addition of the noa-essential amino acids. The RTB, A549, and A375 cells were propagated in Dulbecco's modified MEM with glucose increased to 4.5 mg/ml and containing 10 percent FBS. W a y m o u t h MB 752/1 medium mixed with an equal volume of Gey's salt solution containing FBS was used to grow MEW, CaSki and Dot cells (10 percent FBS for H E W and 20 percent for the other lines). RPMI-1640 medium was used for the growth of other cells, supplemented with 10 percent FBS for 1N!C-37 and HT-1376 cells and 20 percent FBS for lgT-1417 and 8392B cells.
Viruses Bluetongue virus (BTV) serotype 10 (20), purchased from Colorado Serum Co. (Denver, CO) as a vaccine strain designated BT-8, was propagated in BHK-21 cells as described previously (24). Inactivation by ultraviolet (UV) irradiation was described previously (24). Sendal virus (NIX V 321-001-000) was propagated in chicken embryos and titrated in primary hamster kidney cell monolayers (16). Newcastle disease virus (NDV), CG strain, was prepared and titrated as described previously (30).
Inter]eron Induction Induction of interferon in lymphoblastoid cell lines followed the standard procedure described by CANTELL (6) for leukocyte cultures except t h a t we used RPML1640 with 5 percent serum. Other cells were induced by attaching virus to washed confluent monolayers of cells grown for approximately 4 to 6 days in 15× 125 m m screw-cap culture tubes, 30 ml plastic flasks, or 32 oz prescription bottles. At the end of the attachment period, a medium appropriate for the cell was added; the concentration of FBS was reduced to 2 percent, and the volmne was decreased to half that used for cell growth. For cells propagated in RPMI-1640 medium, McCoy's 5a or MEM supplemented with 2 percent FBS was used for interferon production. Supernatant fluids were collected 24 hours after the virus inoculation and stored at --70 ° C. Before titration of interferon, residual inducer virus was acid-inactivated as described previously (23).Polyriboinosinate. p o l y r i b o c y t i d y l a t e - - p o l y ( I ) , p o l y ( C ) - - ( P - L Biochemicals, Milwaukee, WI) was used to induce interferon in the manner described earlier (23).
Inter]etch Assay Interferon was quantitated by a hemagglutinin (HA) yield-reduction assay using encephalomyocarditis (EMC) virus (22) with the BUD-8 cell strain (37). Over the five-year period in which the screening of cell lines was conducted, the sensitivity of the interferon assay increased approximately three-fold. During the initial three years (n = 19) the assay measured about seven times the titer assigned to the I~TIH reference leukocyte interferon (G 023-901-527) ; in 1977 (n = 5) the assay measured about 20 times the titer assigned to the standard. The mean ratio for all titrations (n = 24, GMT i05-26 units/ml, SD = 10 °.3) of the standard suggests that, overall, this assay detects roughly 9 times its assigned value. The EMC HA yield-reduction assay measured about 15 Arch. Virol. 621~
212
PATRIClA J~.~ESO~ and S. E. G~oss]3E~G:
seven times the assigned potency of the N I H reference fibroblast interferon (G023-902527); a GMT of 68,000 units/ml was obtained in 15 titrations (SD = 1.7-fold). The relative sensitivity esti~nated from the ratio of observed/assigned titers of the fibroblast interferon standard was used for the adjustment of mean titers calculated from large numbers of samples, known to contain predominantly fibroblast-like interferon, to express their activities in terms of reference units (RU). A leukocyte interferon preparation used as a laboratory reference was included in every titration to detect assay v~ia bility. In addition to the observed difference in sensitivity of the assay, there is also a difference in the slopes of the dose response curves of leukocyte and fibroblast interferons as previously reported by ED¥ et al. (7), and confirmed in our own assay; in five titrations of the two NIH reference reagents the difference between slopes was small but highly significant (P ~ i0-3). Because of these differences in biological activities of the two types of interferons, and the fact that for most of the cells we tested, no determination was made as to the type of interferon produced, no adjustment of observed titers was made. Further, recent reports (18) indicate that any given cell type may produce leukocyte- and/or fibroblast-like interferons in varying proportions peculiar to the given cell.
Results
Inter/eron-Producing Ability o/ Diverse Cell Lines Because cells m a y v a r y in their capacity to produce interferon in relation to the inducer used, the interferon-producing ability of several cell lines was tested with one or more of three different inducers (Table 1): p o l y ( I ) , poly(C) complexed with DEAE-dextran, NDV, and BTV. The cell lines were derived from tumors or from presumably normal tissues intentionally infected with a transforming virus in vitro. All but two of the cell lines produced at least small amounts of interferon in response to at least one of the viral inducers. KC cells responded to p o l y ( I ) , poly(C) but not to NDV. The response noted for HuF/A28 is not large enough to distinguish it from the antiviral activity of diluted poly (I) • poly (C) t h a t might be carried over from the induced cell culture (see Table 1). Several lines responded well to only one of the inducers. The NC-37 line provided the best response of the three 15~nphoblastoid cells we tested. Another seven cell lines (iN:B, A549, A375, CaSki, Dot, H E W , and HT-1376) produced good yields of interferon ( ~ 1000 units/ml) in response to one or both of the viral inducers; four of these, three from carcinomas, produced more t h a n 6000 units/ml with a viral inducer. Among the lines which produced good yields of interferon, the IIT-1376 cell was the most easily propagated. Two cell lines derived from carcinomas (Dot and H E W ) produced especially high yields of interferon in response to BTV. However, these tumor cell lines grow very slowly and are difficult to propagate so t h a t their interferon-producing potential was not. studied further. Additional experiments on BTV induction of interferon in A375 cells resulted in interferon yields similar to those obtained in the initial screening tests (2000 to 5200 units/ml). The interferon production b y A375 cells remained constant with BTV multiplicities of 1.0, 10.0, or 100 PFU/cell, with which interferon yields of 5200, 3700, and 3700 units/mI, respectively, were obtained. However, UV-inactivated BTV (BTV-UV) at a multiplicity of 50 particles/cell induced only small amounts of interferon in A375 cells with a m a x i m u m of 900 units/ml.
H u m a n Tumor Cell Lines Produce Interferon
213
Table 1. InterJeron production by human cell linesa and strains Yield of interferon (units/ml) with the indicated inducer c Induced cell b
BTV
NDV
poly (I) • poly (C)
Normal cells BUD-8 skin strain t t E G embryo glial strain
50 nd
720 370
HeLa eeU derivatives F L amnion H E K embryo kidney
nd nd
68 75
nd 680
Virus in/ected cells *KC neuroblastoma astrocytes
nd
< 10
440
with integrated I~SV genome HuF/A 28 SV 40 transformed skin
nd a nd
~ I0
~ I0
15
(Faneoni anemia) NB SV40 transformed kidney
2,500
2,500
15
Lymphoblastoid cells *HT-1417 lymphoblastoid "8392B Iymphoblastoid NC-37 lymphoblastoid
25 200 1,200
25 15 1,200
170 210 970
Tumor derived cells *A549 lung carcinoma *A375 melanoma *CaSki cervical carcinoma *Dot cervical carcinoma
480 6,600 2,t00 7,000
2,500 200 2,100 3,000
15 15 nd nd
15,000 4,000
3,000 9,500
nd 1,600
*HEW ovarian carcinoma *HT-1376 bladder carcinoma
Cell lines derived from human tumors are designated by an asterisk b Each milliliter of interferon is the product of about 106 cells c Inducer virus doses: 20 to 50 PFU/cell. Poly (I) • poly (C), complexed with an equal weight or 10-fold excess of DEAE-dextran, was applied to cells at a concentration of i00 ~g/ml a Not done. Interferon induction by poly (I) • poly (C) was not tested in BUD-8 cells as done for other cells. To measure the antiviral effect which might be observed by simple transfer of the poly (I) • poly (C) used as an inducer in various cell cultures, BUD-8 cells were exposed for 24 hours to poly (I) • poly (C) dilutions in medium from I0 i~g/ml through 0.1 ~g/ml. An apparent antiviral titer of 2 was estimated from this control test to be possibly due to residual inducer.
Inter/eron Production by H T - 1 3 7 6 Cells A m o n g the cells we screened, the I:IT-1376 bladder carcinoma celt line appeared to be the most promising cell line for the p r o d u c t i o n of h u m a n interferon i n large q u a n t i t y . The cultures m a y be s u b d i v i d e d weekly a t a ratio of t : 4, reaching a cell d e n s i t y of a b o u t 2 × 104/cm u. I n t e r f e r o n i n d u c t i o n using conditions defined previously (23) resulted i n a geometric m e a n t i t e r (GMT) of 2600 R U / m l (GMT ---15"
214
PATI~ICIA JANESON
and S. E. G~OSSBERG:
18,500 units/ml in a 6-fold range of 10,000 to 62,000 units/ml) from 22 preparations obtained with cultures of about 2 × 107 cells in 32 oz bottles induced with B T V at a multiplicity of about 10 PFU/cell. Subsequently, a series of 37 inductions with a multiplicity of about 1 PFU/cell yielded a GMT of 2800 g U / m l (range of 6300 to 38,000 units/mt). I n the cumulative results of 59 inductions, the interferon yield from HT-1376 celt cultures approached 2700 t~U/ml or a p p r o x i m a t e l y 3 RU/103 cells. HT-1376 cells produced as m u c h interferon when induced with B T V - U V as with infectious B T V (Fig. 1); a multiplicity of 10 PFU/eell was used for the infectious virus, and the same particle dose was used for the inactivated BTV. Virus preparations irradiated until no infectivity was detectable still caused some cytopathie effect (CPE) ; further irradiation reduced the C P E to a negligible level without impairing the interferon-inducing capacity.
I00 z5 L) 50 ,~..4, 25
C
II!i,
Io 5
lo 4
~_ lOs
_--
8
E
\
o~ ~,.
103
106
"I-
.c_ •~
102 ;
104`
>
01
102
"C c
