Proc. Natl. Acad. Sci. USA Vol. 73, No. 12, pp. 4645-4648, December 1976
Immunology
Human alpha-fetoprotein as a modulator of human lymphocyte transformation: Correlation of biological potency with electrophoretic variants (hepatoma/glycoprotein/sialic acid/phytomitogens/fetal development)
ERIC P. LESTER, J. BRUCE MILLER, AND STANLEY YACHNIN Department of Medicine, The University of Chicago, and The Franklin McLean Memorial Research Institute, Chicago, Illinois 60637
Communicated by Leon 0. Jacobson, September 27, 1976
Human alpha-fetoprotein (HAFP) isolated by ABSTRACT immunoadsorbent column was shown to suppress the mitogenic response of human lymphocytes to phytomitogens, antihuman thymocyte antiserum, and the mixed lymphocyte culture. HAFP isolated from the sera and ascitic fluid of five hepatoma patients, and from fetal liver, varied in biological potency over three orders of magnitude. Extended agarose gel electrophoresis and crossed immunoelectrophoresis demonstrated three molecular species of HAFP. Quantitation of the three species revealed a correlation between the relative amount of the most negatively charged species and biological potency. Treatment of HAFP with neuraminidase to remove completely sialic acid residues did not alter the biological potency, but converted the three species to two species having slower electrophoretic mobilities. We conclude that differences in sialic acid content are only partly responsible for the microheterogeneity demonstrated by HAFP, and that variability in another charged moiety is also present. Variation in the relative proportions of the different molecular species of HAFP may be important in the regulation of its immunosuppressive properties. While characterizing the immunosuppressive effects of human alpha-fetoprotein (HAFP) (1), we noted that HAFP isolates from the serum and ascitic fluid of five patients with hepatoma or from fetal sources varied over three orders of magnitude in their capacity to inhibit in vitro human lymphocyte transformation induced by phytomitogens, antihuman thymocyte antiserum (ATS), and the mixed lymphocyte culture (MLC) (2). The most potent HAFP preparation was that isolated from fetal liver homogenates of stillborn human abortuses of 10-20 weeks' gestation (2). Because HAFP isolates display a microheterogeneity with respect to charge (3, 4), presumably due to differences in sialic acid content, we undertook a study of the effects of such charge differences upon the biological potency of our various preparations. The results indicate that the most negatively charged species of HAFP are those with the greatest capability to suppress the mitogenic responses of human lymphocytes in vitro. We also observed that total desialylation of HAFP preparations does not affect their biological potency, nor does it abolish their microheterogeneity.
MATERIALS AND METHODS Human alpha-fetoprotein was isolated from sera and ascitic fluid of five hepatoma patients (Od., McF., Ho., Cr., and Lu.) by a modification of the method of Hirai et al. (1, 2, 5). After elution from the immunoadsorbent column, and prior to passage over Sephadex G-150, the HAFP eluate was passed over a second immunoadsorbent column to which the Na2SO4precipitated globulin fraction of rabbit antihuman serum Abbreviations: HAFP, human alpha-fetoprotein; MLC, mixed lymphocyte culture; ATS, rabbit antiserum against human thymocytes; Con A, concanavalin A. 4645
antibody was attached. HAFP was also isolated from the livers of still-born human abortuses of 10-20 weeks' gestation. The livers were homogenized in phosphate-buffered saline at pH 7.4, centrifuged at 20,000 X g for 1 hr (40), and the soluble liver extract was subjected to the procedures described above. The techniques for the isolation of peripheral blood lymphocytes from normal human adults, as well as methods for their culture, have been described previously (6, 7). The potency of HAFP preparations in suppressing human lymphocyte transformation was derived from dose-response curves plotted on semilogarithmic graph paper and is expressed as the concentration of HAFP required to produce 50% inhibition. The following concentrations of mitogen were employed: Concanavalin A (Con A), 100 ,g/ml; phytohemagglutinin, 10 ,ug/ml; pokeweed mitogen, 0.08,ug/ml; ATS, 25,ul/ml. Sialic acid quantitations were done by a modification of the method of Warren (8, 9). After exhaustive dialysis against distilled water at 40, HAFP preparations were lyophilized and acid desialylation was performed in 0.03 M HCl at 800 for 45 min. Enzymatic desialylation was carried out with a neuraminidase derived from Clostridium perfringens (Sigma Chemical Co., St. Louis, Mo.). HAFP (2 mg/ml) was digested with 0.06 units of enzyme per ml at 370 in phosphate-buffered saline at pH 7.0) for periods of up to 18 hr. Under these conditions, digestion was complete by 3 hr, as monitored by measurement of the release of free sialic acid. After neuraminidase digestion, portions of the desialylated HAFP were heated at 600 for 1 hr to inactivate residual neuraminidase, a procedure which was shown not to affect the electrophoretic mobility or biological activity of native HAFP (2). Similar results were achieved with a neuraminidase derived from Vibrio cholera (Behring Diagnostics, Somerville, N.J.), and with a Sepharose-bound neuraminidase from Clostridium perfringens (Sigma Chemical Co., St. Louis, Mo.). Two-dimensional crossed immunoelectrophoresis was performed with 1% agarose [Seakem (LE), Marine Colloids, Rockland, Me.] at pH 8.6, in 0.15 M barbital buffer containing 0.003 M calcium lactate (4). The proportions of the various HAFP electrophoretic variants in any HAFP isolate were estimated by cutting out and weighing the estimated area under each peak from paper tracings of the crossed immunoelectrophoretic patterns. In addition, after extended agarose gel electrophoresis at pH 8.6, certain plates were fixed, dried, stained with Amido black, and scanned at 560 nm with a linear transport chromatogram spectrophotometer (Zeiss PMQ II) for quantitation of the various HAFP electrophoretic variants. RESULTS To explain the variability in the biological potency of various HAFP preparations, we first examined their sialic acid content,
Lester et al.
Immunology:
4646
Proc. Natl. Acad. Sci. USA 73 (1976)
Table 1. Biological potency, sialic acid content, and molecular species of HAFP Biological potency
(0g/ml)*
HAFP source Fetal liver Hepatoma Hepatoma Hepatoma Hepatoma Hepatoma
patient patient patient patient patient
3 20 130 500 1000 >2000
Lu. Cr. Ho. McF. Od.
% of total HAFP
Moles sialic acid per mole HAFP
1.39t 1.53 + 1.40 + 1.54 + 1.13 + 1.45 +
0.17 0.16 0.20 0.19 0.12
HAFP 1
HAFP 2
19.9 19.8 27.2 29.7 27.9
50.3 50.9 44.7 41.1 53.3
HAFP 3 HAFP 3:HAFP 1 100 29.8 29.3 28.1 29.3 18.9
1.49 1.48 1.05 0.99 0.69
* Amount of HAFP producing 50% inhibition of lymphocyte mitogenic response to ATS, non-hemagglutinating phytohemagglutinin, Con A, or pokeweed mitogen. (See ref. 2.) t Single determination. All other determinations represent the mean 3-6 experiments 4 SID
because desialylation of glycoprotein hormones was shown in some cases to increase their biological potency in vitro (10, 11). As can be seen in Table 1, no correlation was found between sialic acid content and the biological potency of the various HAFP isolates. In addition, complete enzymatic desialylation of our two most active HAFP preparations did not alter their capacity to inhibit human lymphocyte transformation in vitro (Table 2). Because previous workers had noted microheterogeneity of HAFP by means of crossed immunoelectrophoresis (3, 4), we applied this technique to a study of our purified HAFP preparations, and of the sera from which they were isolated. When concentrated solutions of purified hepatoma HAFP (10 mg/ml) were examined by one-dimensional agarose gel electrophoresis, three bands, the intensity of which varied from one isolate to another, were easily discerned (Fig. 1). The most cathodemigrating band was termed HAFP 1, the middle band HAFP 2, and the most anode-migrating band HAFP 3. The RF values for these bands relative to the cathodal margin of serum albumin were 0.911, 0.944, and 0.975. Analysis of these HAFP isolates by crossed immunoelectrophoresis confirmed the
of three HAFP peaks in each of the hepatoma HAFP isolates studied, albeit in varying proportions (Fig. 2). The relative proportions of the three peaks in a given isolate were constant when the sample was quantitated by crossed immunoelectrophoresis and by spectrophotometric scanning. In each of three instances in which fresh-frozen (-85°) serum was available for study, similar heterogeneity of HAFP could be discerned (Fig. 2F). When the hepatoma HAFP isolates were completely desialylated by neuraminidase digestion, two HAFP peaks were seen consistently; the most anode-migrating one now had an electrophoretic mobility corresponding to the most cathode-migrating peak previously seen in the native HAFP isolate (Fig. 2F). The most cathode-migrating of the desialylated HAFP peaks was termed HAFP 0. Analysis of fetal liver HAFP (800 ,ug/ml) by similar techniques revealed the presence of a single peak whose mobility corresponded to the most anode-migrating one of the native hepatoma HAFP species. Desialylated fetal liver HAFP also revealed a single peak, which was now shifted to the position of the most cathodal native hepatoma HAFP band (Fig. 2F). The similarity of the differences in RF between neighboring
presence
Table 2. Effect of desialylation upon the capacity of HAFP to inhibit human lymphocyte transformation [HAFP]
HAFP
Mitogen
(mg/ml)
None (control) Patient Lu., native
Con A Con A
0 50 20 10 50 20 10 6.4 3.2 6.4 3.2 0 50 20 10 50 20 10
Patient Lu.,
Con A
desialylated Fetal liver, native Fetal liver,
desialylated None (control) Patient Lu.,
Con A Con A ATS ATS
native Patient Lu.,
desialylated
ATS
HAFP
[2-' 4C ] thymidine incorporation (cpm ± SD)
Inhibition (%)
65,016 + 2,472 17,013 + 959 38,976 + 2,249
73.8 40.0
54,397 ± 5,489 18,291 ± 160 41,947 ± 4,439 54,916+ 209 20,122 + 634 46,739 ± 4,383 26,412 + 1,048 46,997 ± 1,011 68,271 + 5,550 21,572 + 45 47,547 ± 2,636 56,442 ± 431 22,689 + 969 46,460 + 2,032 52,555 + 991
* Amount of HAFP producing 50% inhibition of lymphocyte transformation.
71.9 35.5 15.5 69.0 28.1 59.4 27.7 68.4 30.4 17.3 66.8 31.9 23.0
biological potency* (g/ml)
26
28.5 4.7 5.2
32
32
Immunology:
a
Lester et al.
a.
#
Proc. Natl. Acad. Sci. USA 73 (1976)
4647
;9
I
b I*
-c
+
d e
I
FIG. 1. Extended agarose gel electrophoresis of HAFP. (a) Normal human serum; (b) HAFP (patient Cr.); (c) HAFP (patient Ho.); (d) HAFP (patient Od.); (e) normal human serum.
pairs of the four HAFP bands encountered both in native and desialylated HAFP preparations suggests that a unit charge difference accounts for each decrement in electrophoretic mobility (Table 3). Because our biologically most active HAFP preparation, fetal liver HAFP, consisted almost exclusively of HAFP 3, we correlated the relative biological potencies of the hepatoma HAFP isolates with their proportional content of HAFP 3, and with the ratio of HAFP 3:HAFP 1. As shown in Table 1, those HAFP isolates containing the highest HAFP 3:HAFP 1 ratios had the highest biological activity.
DISCUSSION Previous workers studied the microheterogeneity of HAFP by a variety of techniques that included starch gel electrophoresis at pH 5.0 (3), isoelectric focusing (4), ion exchange chromatography (4), and crossed immunoelectrophoresis (3, 4). Two major variants were identified; Purves, however, encountered a hepatoma serum displaying three HAFP electrophoretic variants (4), and Alpert et al. described additional HAFP microheterogeneity by both isoelectric focusing and crossed immunoelectrophoresis (4). We also observed additional HAFP microheterogeneity by isoelectric focusing. * Several lines of evidence indicate that these electrophoretic variants are not artefacts of the method of preparation, because (i) the variants are seen in native serum samples which are freshly drawn, or freshly frozen at -85° and (ii) they are present in HAFP isolates prepared by a variety of methods that include immunoadsorbent columns, isoelectric focusing, and ion exchange chromatography (3, 4). Neither can they be explained by simple variability in the HAFP sialic acid content, because we have shown that the isoelectric peaks of HAFP at pH 4.8 and 5.2 each contain the same proportion of sialic acid as the HAFP isolate from which they are prepared.t In addition, if sialic acid variability were the only basis for the microheterogeneity, then complete desialylation of HAFP, which we documented, would be expected to result in a single species of *S. Yachnin, R. Hsu, R. L. Heinrikson, and J. B. Miller. Studies on human alpha-fetoprotein: Isolation and characterization of monomeric and multimeric forms, and amino-terminal sequence analysis. Submitted for publication. tJ. B. Miller, E. P. Lester, and S. Yachnin. Unpublished observation.
HAFP only, whereas we consistently see two residual HAFP peaks in totally desialylated hepatoma HAFP. If each HAFP molecule contains one or two sialic acid molecules, as seems likely from our sialic acid analyses, and if, in addition, the HAFP molecules are heterogeneous with respect to another charged moiety, then the presence of three HAFP species in native HAFP isolates, and of two in the desialylated preparations, is readily explained. Whether this other charged moiety is a genetically determined amino acid substitution, a post-synthetic physiological or artefactual change in chemical composition (e.g., deamidation), or a conformation change, remains to be determined. Whatever the mechanism by which it arises, the structural basis of this second charge difference, rather than sialic acid heterogeneity, may determine how effectively HAFP isolates inhibit human lymphocyte transformation. Thus, the fetal liver HAFP isolate, which consists exclusively or predominantly of HAFP 3, is the most potent lymphocyte inhibitor we have encountered, and desialylation, with a concomitant shift in mobility corresponding to two unit charges, does not alter its biological potency. In addition, the biological potency of hepatoma HAFP isolates can be correlated with the relative amounts of HAFP 3 which they contain. Thus, HAFP, which serves as an immunoregulator during fetal life, may be subject to both qualitative and quantitative controls, during fetal development, that serve to temper the effectiveness with which it modulates lymphocyte responsiveness. It is conceivable that immunosuppression due to HAFP not only diminishes as fetal HAFP synthesis and serum concentration subside during fetal development, but that, at critical periods during fetal life, genetic controls dictate the production of more or less potent immunosuppressive molecular species of HAFP. Reappearance of potent immunosuppressive species of HAFP in some hepatoma patients may, in contrast, be viewed as a more random type of gene activation occurring during neoplastic transformation. Table 3. RF values of HAFP variants
HAFPO HAFP1 HAFP2 HAFP3
*
-
0.911 0.920 0.899
0.944 0.949
0.975 0.986
(-* )*
(-2)*
(-3)*
Purified HAFP Serum HAFP Desialylated HAFP
0.867
Presumed charge relative to HAFP 0.
Immunology:
4648
1.0
Lester et al.
Proc. Natl. Acad. Sci. USA 73 (1976)
-__
B) McF
A) Od. HAFP 2
HAFP
C) Ho.
HAFP 2 HAFP .- 3
A560 nm
-
HAFP 3
+
0 1.0
-
D) Cr
F)
E)Lu
A
Purified HAFP (Od.) Serum HAFP (Od.)
A560nm
Desial. HAFP (Od.) +--i
_ F \ < / v /lisP+
0
-
,'"..
_-
Fetal Liver HAFP Desial. Fetal Liver HAFP _
+
FIG. 2. Molecular species of HAFP. Panels A-E: the solid lines (-) represent the spectrophotometric scans, at 560 nm, of extended agarose gel electrophoreses of HAFP isolates prepared from individual hepatoma patient sera. Beneath each of these is displayed the result of crossed immunoelectrophoresis as well as the Amido black-stained samples from extended agarose gel electrophoresis of the same HAFP isolate. The latter two are reproduced to the same scale. In all instances, three electrophoretic species of HAFP, designated HAFP 1, HAFP 2, and HAFP 3 are seen. The anode is to the right. Panel F: crossed immunoelectrophoretic patterns of HAFP. The crossed immunoelectrophoretic pattern of HAFP (patient Od.) is identical in the serum and the purified HAFP isolate. After desialylation (desial.), HAFP 2 and HAFP 3 are no longer visible, and a new HAFP species (HAFP 0) has appeared. Fetal liver HAFP consists exclusively of HAFP 3 and is converted by desialylation to HAFP 1.
This investigation was supported by a National Institutes of Health Fellowship (no.5 F32 CA05337-02) from the National Cancer Institute, by grant number CA 14599-03 awarded by the National Cancer Institute, Department of Health, Education, and Welfare, and by a grant from the Leukemia Research Foundation. The Franklin McLean Memorial Research Institute is operated by the University of Chicago for the U.S. Energy Research and Development Administration under Contract E(11-1)-69). 1. Yachnin, S. (1976) "Demonstration of the inhibitory effects of human alpha-fetoprotein on in vitro transformation of human lymphocytes," Proc. Natl. Acad. Sci. USA 73,2857-2861. 2. Yachnin, S. & Lester, E. P. (1976) "Inhibition of human lym-
phocyte transformation by human alpha-fetoprotein (HAFP); comparison of fetal and hepatoma HAFP and kinetic studies of in vitro immunosuppression," Clin. Exp. Immunol., in press. 3. Purves, L. R., Van der Merwe, E. & Bersohn, I. (1970) "Serum alpha-fetoprotein. V. The bulk preparation and some properties of alpha-fetoprotein obtained from patients with primary cancer of the liver," S. Afr. Med. J. 44, 1264-1268. 4. Alpert, E., Drysdale, J. W., Isselbacher, K. J. & Schur, P. H. (1972) "Human alpha-fetoprotein: Isolation, characterization, and demonstration of microheterogeneity," J. Biol. Chem. 247, 3792-3798. 5. Hirai, H., Nishi, S., Watabe, H. & Tsukada, Y. (1973) "Alphafetoprotein and cancer," Gann Monograph on Cancer Research, eds. Hirai, H. & Miyayi, T. (University of Tokyo Press, Tokyo), pp. 19-34.
6. Allen, L. W., Svenson, R. H. & Yachnin, S. (1969) "Purification of mitogenic proteins derived from Phaseolus vulgaris. Isolation of potent and weak phytohemagglutinins," Proc. Natl. Acad. Sci. USA 63, 334-341.
7. Yachnin, S. (1972) "The potentiation and inhibition by autologous red cells and platelets of human lymphocyte transformation induced by pokeweed mitogen, concanavalin A, mercuric chloride, antigen, and mixed leucocyte culture," Clin. Exp. Immunol. 11, 109-124. 8. Hahn, H.-J., Hellman, B., Lernmark, A., Sehlin, J. & Taljedal, I.-B. (1974) "The pancreatic beta-cell recognition of insulin secretagogues. The influence of neuraminidase treatment on the release of insulin and the islet content of insulin, sialic acid, and cyclic adenosine 3',5'-monophosphate," J. Biol. Chem. 249, 5275-5284.
9. Warren, L. (1959) "The thiobarbituric acid assay of sialic acids," J. Biol. Chem. 234, 1971-1975. 10. Goldwasser, E., Kung, C. & Eliason, J. (1974) "On the mechanism of erythropoietin-induced differentiation. XIII. The role of sialic acid in erythropoietin action," J. Biol. Chem. 249, 4202-4206. 11. Dufau, M. L., Catt, K. H. & Tsuruhara, T. (1971) "Retention of in vitro biological activities by desialylated human luteinizing hormone and chorionic gonadotropin," Biochem. Biophys. Res. Commun. 44, 1022-1029.
Immunology
Human alpha-fetoprotein as a modulator of human lymphocyte transformation: Correlation of biological potency with electrophoretic variants (hepatoma/glycoprotein/sialic acid/phytomitogens/fetal development)
ERIC P. LESTER, J. BRUCE MILLER, AND STANLEY YACHNIN Department of Medicine, The University of Chicago, and The Franklin McLean Memorial Research Institute, Chicago, Illinois 60637
Communicated by Leon 0. Jacobson, September 27, 1976
Human alpha-fetoprotein (HAFP) isolated by ABSTRACT immunoadsorbent column was shown to suppress the mitogenic response of human lymphocytes to phytomitogens, antihuman thymocyte antiserum, and the mixed lymphocyte culture. HAFP isolated from the sera and ascitic fluid of five hepatoma patients, and from fetal liver, varied in biological potency over three orders of magnitude. Extended agarose gel electrophoresis and crossed immunoelectrophoresis demonstrated three molecular species of HAFP. Quantitation of the three species revealed a correlation between the relative amount of the most negatively charged species and biological potency. Treatment of HAFP with neuraminidase to remove completely sialic acid residues did not alter the biological potency, but converted the three species to two species having slower electrophoretic mobilities. We conclude that differences in sialic acid content are only partly responsible for the microheterogeneity demonstrated by HAFP, and that variability in another charged moiety is also present. Variation in the relative proportions of the different molecular species of HAFP may be important in the regulation of its immunosuppressive properties. While characterizing the immunosuppressive effects of human alpha-fetoprotein (HAFP) (1), we noted that HAFP isolates from the serum and ascitic fluid of five patients with hepatoma or from fetal sources varied over three orders of magnitude in their capacity to inhibit in vitro human lymphocyte transformation induced by phytomitogens, antihuman thymocyte antiserum (ATS), and the mixed lymphocyte culture (MLC) (2). The most potent HAFP preparation was that isolated from fetal liver homogenates of stillborn human abortuses of 10-20 weeks' gestation (2). Because HAFP isolates display a microheterogeneity with respect to charge (3, 4), presumably due to differences in sialic acid content, we undertook a study of the effects of such charge differences upon the biological potency of our various preparations. The results indicate that the most negatively charged species of HAFP are those with the greatest capability to suppress the mitogenic responses of human lymphocytes in vitro. We also observed that total desialylation of HAFP preparations does not affect their biological potency, nor does it abolish their microheterogeneity.
MATERIALS AND METHODS Human alpha-fetoprotein was isolated from sera and ascitic fluid of five hepatoma patients (Od., McF., Ho., Cr., and Lu.) by a modification of the method of Hirai et al. (1, 2, 5). After elution from the immunoadsorbent column, and prior to passage over Sephadex G-150, the HAFP eluate was passed over a second immunoadsorbent column to which the Na2SO4precipitated globulin fraction of rabbit antihuman serum Abbreviations: HAFP, human alpha-fetoprotein; MLC, mixed lymphocyte culture; ATS, rabbit antiserum against human thymocytes; Con A, concanavalin A. 4645
antibody was attached. HAFP was also isolated from the livers of still-born human abortuses of 10-20 weeks' gestation. The livers were homogenized in phosphate-buffered saline at pH 7.4, centrifuged at 20,000 X g for 1 hr (40), and the soluble liver extract was subjected to the procedures described above. The techniques for the isolation of peripheral blood lymphocytes from normal human adults, as well as methods for their culture, have been described previously (6, 7). The potency of HAFP preparations in suppressing human lymphocyte transformation was derived from dose-response curves plotted on semilogarithmic graph paper and is expressed as the concentration of HAFP required to produce 50% inhibition. The following concentrations of mitogen were employed: Concanavalin A (Con A), 100 ,g/ml; phytohemagglutinin, 10 ,ug/ml; pokeweed mitogen, 0.08,ug/ml; ATS, 25,ul/ml. Sialic acid quantitations were done by a modification of the method of Warren (8, 9). After exhaustive dialysis against distilled water at 40, HAFP preparations were lyophilized and acid desialylation was performed in 0.03 M HCl at 800 for 45 min. Enzymatic desialylation was carried out with a neuraminidase derived from Clostridium perfringens (Sigma Chemical Co., St. Louis, Mo.). HAFP (2 mg/ml) was digested with 0.06 units of enzyme per ml at 370 in phosphate-buffered saline at pH 7.0) for periods of up to 18 hr. Under these conditions, digestion was complete by 3 hr, as monitored by measurement of the release of free sialic acid. After neuraminidase digestion, portions of the desialylated HAFP were heated at 600 for 1 hr to inactivate residual neuraminidase, a procedure which was shown not to affect the electrophoretic mobility or biological activity of native HAFP (2). Similar results were achieved with a neuraminidase derived from Vibrio cholera (Behring Diagnostics, Somerville, N.J.), and with a Sepharose-bound neuraminidase from Clostridium perfringens (Sigma Chemical Co., St. Louis, Mo.). Two-dimensional crossed immunoelectrophoresis was performed with 1% agarose [Seakem (LE), Marine Colloids, Rockland, Me.] at pH 8.6, in 0.15 M barbital buffer containing 0.003 M calcium lactate (4). The proportions of the various HAFP electrophoretic variants in any HAFP isolate were estimated by cutting out and weighing the estimated area under each peak from paper tracings of the crossed immunoelectrophoretic patterns. In addition, after extended agarose gel electrophoresis at pH 8.6, certain plates were fixed, dried, stained with Amido black, and scanned at 560 nm with a linear transport chromatogram spectrophotometer (Zeiss PMQ II) for quantitation of the various HAFP electrophoretic variants. RESULTS To explain the variability in the biological potency of various HAFP preparations, we first examined their sialic acid content,
Lester et al.
Immunology:
4646
Proc. Natl. Acad. Sci. USA 73 (1976)
Table 1. Biological potency, sialic acid content, and molecular species of HAFP Biological potency
(0g/ml)*
HAFP source Fetal liver Hepatoma Hepatoma Hepatoma Hepatoma Hepatoma
patient patient patient patient patient
3 20 130 500 1000 >2000
Lu. Cr. Ho. McF. Od.
% of total HAFP
Moles sialic acid per mole HAFP
1.39t 1.53 + 1.40 + 1.54 + 1.13 + 1.45 +
0.17 0.16 0.20 0.19 0.12
HAFP 1
HAFP 2
19.9 19.8 27.2 29.7 27.9
50.3 50.9 44.7 41.1 53.3
HAFP 3 HAFP 3:HAFP 1 100 29.8 29.3 28.1 29.3 18.9
1.49 1.48 1.05 0.99 0.69
* Amount of HAFP producing 50% inhibition of lymphocyte mitogenic response to ATS, non-hemagglutinating phytohemagglutinin, Con A, or pokeweed mitogen. (See ref. 2.) t Single determination. All other determinations represent the mean 3-6 experiments 4 SID
because desialylation of glycoprotein hormones was shown in some cases to increase their biological potency in vitro (10, 11). As can be seen in Table 1, no correlation was found between sialic acid content and the biological potency of the various HAFP isolates. In addition, complete enzymatic desialylation of our two most active HAFP preparations did not alter their capacity to inhibit human lymphocyte transformation in vitro (Table 2). Because previous workers had noted microheterogeneity of HAFP by means of crossed immunoelectrophoresis (3, 4), we applied this technique to a study of our purified HAFP preparations, and of the sera from which they were isolated. When concentrated solutions of purified hepatoma HAFP (10 mg/ml) were examined by one-dimensional agarose gel electrophoresis, three bands, the intensity of which varied from one isolate to another, were easily discerned (Fig. 1). The most cathodemigrating band was termed HAFP 1, the middle band HAFP 2, and the most anode-migrating band HAFP 3. The RF values for these bands relative to the cathodal margin of serum albumin were 0.911, 0.944, and 0.975. Analysis of these HAFP isolates by crossed immunoelectrophoresis confirmed the
of three HAFP peaks in each of the hepatoma HAFP isolates studied, albeit in varying proportions (Fig. 2). The relative proportions of the three peaks in a given isolate were constant when the sample was quantitated by crossed immunoelectrophoresis and by spectrophotometric scanning. In each of three instances in which fresh-frozen (-85°) serum was available for study, similar heterogeneity of HAFP could be discerned (Fig. 2F). When the hepatoma HAFP isolates were completely desialylated by neuraminidase digestion, two HAFP peaks were seen consistently; the most anode-migrating one now had an electrophoretic mobility corresponding to the most cathode-migrating peak previously seen in the native HAFP isolate (Fig. 2F). The most cathode-migrating of the desialylated HAFP peaks was termed HAFP 0. Analysis of fetal liver HAFP (800 ,ug/ml) by similar techniques revealed the presence of a single peak whose mobility corresponded to the most anode-migrating one of the native hepatoma HAFP species. Desialylated fetal liver HAFP also revealed a single peak, which was now shifted to the position of the most cathodal native hepatoma HAFP band (Fig. 2F). The similarity of the differences in RF between neighboring
presence
Table 2. Effect of desialylation upon the capacity of HAFP to inhibit human lymphocyte transformation [HAFP]
HAFP
Mitogen
(mg/ml)
None (control) Patient Lu., native
Con A Con A
0 50 20 10 50 20 10 6.4 3.2 6.4 3.2 0 50 20 10 50 20 10
Patient Lu.,
Con A
desialylated Fetal liver, native Fetal liver,
desialylated None (control) Patient Lu.,
Con A Con A ATS ATS
native Patient Lu.,
desialylated
ATS
HAFP
[2-' 4C ] thymidine incorporation (cpm ± SD)
Inhibition (%)
65,016 + 2,472 17,013 + 959 38,976 + 2,249
73.8 40.0
54,397 ± 5,489 18,291 ± 160 41,947 ± 4,439 54,916+ 209 20,122 + 634 46,739 ± 4,383 26,412 + 1,048 46,997 ± 1,011 68,271 + 5,550 21,572 + 45 47,547 ± 2,636 56,442 ± 431 22,689 + 969 46,460 + 2,032 52,555 + 991
* Amount of HAFP producing 50% inhibition of lymphocyte transformation.
71.9 35.5 15.5 69.0 28.1 59.4 27.7 68.4 30.4 17.3 66.8 31.9 23.0
biological potency* (g/ml)
26
28.5 4.7 5.2
32
32
Immunology:
a
Lester et al.
a.
#
Proc. Natl. Acad. Sci. USA 73 (1976)
4647
;9
I
b I*
-c
+
d e
I
FIG. 1. Extended agarose gel electrophoresis of HAFP. (a) Normal human serum; (b) HAFP (patient Cr.); (c) HAFP (patient Ho.); (d) HAFP (patient Od.); (e) normal human serum.
pairs of the four HAFP bands encountered both in native and desialylated HAFP preparations suggests that a unit charge difference accounts for each decrement in electrophoretic mobility (Table 3). Because our biologically most active HAFP preparation, fetal liver HAFP, consisted almost exclusively of HAFP 3, we correlated the relative biological potencies of the hepatoma HAFP isolates with their proportional content of HAFP 3, and with the ratio of HAFP 3:HAFP 1. As shown in Table 1, those HAFP isolates containing the highest HAFP 3:HAFP 1 ratios had the highest biological activity.
DISCUSSION Previous workers studied the microheterogeneity of HAFP by a variety of techniques that included starch gel electrophoresis at pH 5.0 (3), isoelectric focusing (4), ion exchange chromatography (4), and crossed immunoelectrophoresis (3, 4). Two major variants were identified; Purves, however, encountered a hepatoma serum displaying three HAFP electrophoretic variants (4), and Alpert et al. described additional HAFP microheterogeneity by both isoelectric focusing and crossed immunoelectrophoresis (4). We also observed additional HAFP microheterogeneity by isoelectric focusing. * Several lines of evidence indicate that these electrophoretic variants are not artefacts of the method of preparation, because (i) the variants are seen in native serum samples which are freshly drawn, or freshly frozen at -85° and (ii) they are present in HAFP isolates prepared by a variety of methods that include immunoadsorbent columns, isoelectric focusing, and ion exchange chromatography (3, 4). Neither can they be explained by simple variability in the HAFP sialic acid content, because we have shown that the isoelectric peaks of HAFP at pH 4.8 and 5.2 each contain the same proportion of sialic acid as the HAFP isolate from which they are prepared.t In addition, if sialic acid variability were the only basis for the microheterogeneity, then complete desialylation of HAFP, which we documented, would be expected to result in a single species of *S. Yachnin, R. Hsu, R. L. Heinrikson, and J. B. Miller. Studies on human alpha-fetoprotein: Isolation and characterization of monomeric and multimeric forms, and amino-terminal sequence analysis. Submitted for publication. tJ. B. Miller, E. P. Lester, and S. Yachnin. Unpublished observation.
HAFP only, whereas we consistently see two residual HAFP peaks in totally desialylated hepatoma HAFP. If each HAFP molecule contains one or two sialic acid molecules, as seems likely from our sialic acid analyses, and if, in addition, the HAFP molecules are heterogeneous with respect to another charged moiety, then the presence of three HAFP species in native HAFP isolates, and of two in the desialylated preparations, is readily explained. Whether this other charged moiety is a genetically determined amino acid substitution, a post-synthetic physiological or artefactual change in chemical composition (e.g., deamidation), or a conformation change, remains to be determined. Whatever the mechanism by which it arises, the structural basis of this second charge difference, rather than sialic acid heterogeneity, may determine how effectively HAFP isolates inhibit human lymphocyte transformation. Thus, the fetal liver HAFP isolate, which consists exclusively or predominantly of HAFP 3, is the most potent lymphocyte inhibitor we have encountered, and desialylation, with a concomitant shift in mobility corresponding to two unit charges, does not alter its biological potency. In addition, the biological potency of hepatoma HAFP isolates can be correlated with the relative amounts of HAFP 3 which they contain. Thus, HAFP, which serves as an immunoregulator during fetal life, may be subject to both qualitative and quantitative controls, during fetal development, that serve to temper the effectiveness with which it modulates lymphocyte responsiveness. It is conceivable that immunosuppression due to HAFP not only diminishes as fetal HAFP synthesis and serum concentration subside during fetal development, but that, at critical periods during fetal life, genetic controls dictate the production of more or less potent immunosuppressive molecular species of HAFP. Reappearance of potent immunosuppressive species of HAFP in some hepatoma patients may, in contrast, be viewed as a more random type of gene activation occurring during neoplastic transformation. Table 3. RF values of HAFP variants
HAFPO HAFP1 HAFP2 HAFP3
*
-
0.911 0.920 0.899
0.944 0.949
0.975 0.986
(-* )*
(-2)*
(-3)*
Purified HAFP Serum HAFP Desialylated HAFP
0.867
Presumed charge relative to HAFP 0.
Immunology:
4648
1.0
Lester et al.
Proc. Natl. Acad. Sci. USA 73 (1976)
-__
B) McF
A) Od. HAFP 2
HAFP
C) Ho.
HAFP 2 HAFP .- 3
A560 nm
-
HAFP 3
+
0 1.0
-
D) Cr
F)
E)Lu
A
Purified HAFP (Od.) Serum HAFP (Od.)
A560nm
Desial. HAFP (Od.) +--i
_ F \ < / v /lisP+
0
-
,'"..
_-
Fetal Liver HAFP Desial. Fetal Liver HAFP _
+
FIG. 2. Molecular species of HAFP. Panels A-E: the solid lines (-) represent the spectrophotometric scans, at 560 nm, of extended agarose gel electrophoreses of HAFP isolates prepared from individual hepatoma patient sera. Beneath each of these is displayed the result of crossed immunoelectrophoresis as well as the Amido black-stained samples from extended agarose gel electrophoresis of the same HAFP isolate. The latter two are reproduced to the same scale. In all instances, three electrophoretic species of HAFP, designated HAFP 1, HAFP 2, and HAFP 3 are seen. The anode is to the right. Panel F: crossed immunoelectrophoretic patterns of HAFP. The crossed immunoelectrophoretic pattern of HAFP (patient Od.) is identical in the serum and the purified HAFP isolate. After desialylation (desial.), HAFP 2 and HAFP 3 are no longer visible, and a new HAFP species (HAFP 0) has appeared. Fetal liver HAFP consists exclusively of HAFP 3 and is converted by desialylation to HAFP 1.
This investigation was supported by a National Institutes of Health Fellowship (no.5 F32 CA05337-02) from the National Cancer Institute, by grant number CA 14599-03 awarded by the National Cancer Institute, Department of Health, Education, and Welfare, and by a grant from the Leukemia Research Foundation. The Franklin McLean Memorial Research Institute is operated by the University of Chicago for the U.S. Energy Research and Development Administration under Contract E(11-1)-69). 1. Yachnin, S. (1976) "Demonstration of the inhibitory effects of human alpha-fetoprotein on in vitro transformation of human lymphocytes," Proc. Natl. Acad. Sci. USA 73,2857-2861. 2. Yachnin, S. & Lester, E. P. (1976) "Inhibition of human lym-
phocyte transformation by human alpha-fetoprotein (HAFP); comparison of fetal and hepatoma HAFP and kinetic studies of in vitro immunosuppression," Clin. Exp. Immunol., in press. 3. Purves, L. R., Van der Merwe, E. & Bersohn, I. (1970) "Serum alpha-fetoprotein. V. The bulk preparation and some properties of alpha-fetoprotein obtained from patients with primary cancer of the liver," S. Afr. Med. J. 44, 1264-1268. 4. Alpert, E., Drysdale, J. W., Isselbacher, K. J. & Schur, P. H. (1972) "Human alpha-fetoprotein: Isolation, characterization, and demonstration of microheterogeneity," J. Biol. Chem. 247, 3792-3798. 5. Hirai, H., Nishi, S., Watabe, H. & Tsukada, Y. (1973) "Alphafetoprotein and cancer," Gann Monograph on Cancer Research, eds. Hirai, H. & Miyayi, T. (University of Tokyo Press, Tokyo), pp. 19-34.
6. Allen, L. W., Svenson, R. H. & Yachnin, S. (1969) "Purification of mitogenic proteins derived from Phaseolus vulgaris. Isolation of potent and weak phytohemagglutinins," Proc. Natl. Acad. Sci. USA 63, 334-341.
7. Yachnin, S. (1972) "The potentiation and inhibition by autologous red cells and platelets of human lymphocyte transformation induced by pokeweed mitogen, concanavalin A, mercuric chloride, antigen, and mixed leucocyte culture," Clin. Exp. Immunol. 11, 109-124. 8. Hahn, H.-J., Hellman, B., Lernmark, A., Sehlin, J. & Taljedal, I.-B. (1974) "The pancreatic beta-cell recognition of insulin secretagogues. The influence of neuraminidase treatment on the release of insulin and the islet content of insulin, sialic acid, and cyclic adenosine 3',5'-monophosphate," J. Biol. Chem. 249, 5275-5284.
9. Warren, L. (1959) "The thiobarbituric acid assay of sialic acids," J. Biol. Chem. 234, 1971-1975. 10. Goldwasser, E., Kung, C. & Eliason, J. (1974) "On the mechanism of erythropoietin-induced differentiation. XIII. The role of sialic acid in erythropoietin action," J. Biol. Chem. 249, 4202-4206. 11. Dufau, M. L., Catt, K. H. & Tsuruhara, T. (1971) "Retention of in vitro biological activities by desialylated human luteinizing hormone and chorionic gonadotropin," Biochem. Biophys. Res. Commun. 44, 1022-1029.
