Britishjournal ofhlaematology, 1g7g,41,343-356.
Vitamin B6 Metabolism in Anaemic and Alcoholic Man L. R. SOLOMON A N D R. S. HILLMAN Department of Medicine, University o f Washington School of Medicine, Seattle, Washington and West Haven Veterans Administration Hospital and Yale University School $Medicine, New Haven, Connecticut (Received 9 November 1977; accepted for publication 28 February 1978) SUMMARY. Physiological and pathological factors affecting intracellular red cell vitamin B6 metabolism in normal, anaemic and alcoholic man were studied using a new assay for pyridoxine kinase (PnK) together with saturated and total aspartate aminotransferase (EGOT) activities as indirect indices of intracellular pyridoxal 5-phosphate (PLP) availability. In studies of anaemic states, subjects with iron deficiency anaemia demonstrated elevated levels of both PnK and saturated EGOT, while seven out of 17 subjects with inflammatory anaemia had subnormal PnK but variable saturated EGOT activities. Despite a high incidence of complicating inflammatory disease, alcoholic subjects with or without ring sideroblastic anaemia had elevated levels of both PnK and saturated EGOT. As judged from the saturated EGOT and the ratio unsaturated EGOT/saturated EGOT, intracellular PLP availability was always appropriate to the higher levels of PnK activity. Hereditary factors, inflammation, neoplasia, exposure to heavy metals, ingestion of drugs and alcoholism have all been implicated in the development of ring sideroblastic anaemias. In alcoholic patients, recent reports have suggested an association with one or more abnormalities in vitamin B6 metabolism, including: low serum levels of the active coenzyme form, pyridoxal 5-phosphate (PLP) (Leevy et al, 1965;Hines& Grasso, 1970; Eichner & Hillman, 1971;Davis & Smith, 1974; Hines & Cowan, 1974); impaired generation of plasma PLP following intravenous administration of pyridoxine (Hines & Cowan, 1970, 1974; Hines, 1975); depressed activity of erythrocyte pyridoxine kinase (PnK), the enzyme which converts the three major dietary forms of vitamin B6 to the active phosphorylated vitamins (Hines, 1975); the appearance of a plasma factor which inhibits PnK activity in vitro (Hines, 1975);and an apparent increase in a specific neutral phosphatase on the red cell membrane capable of dephosphorylating PLP (Lumeng & Li, 1974). The observation that some subjects with alcohol-induced sideroblastic anaemia fail to respond to treatment with pyridoxine but have reticulocyte responses during treatment with PLP also suggests the presence of a block in the intracellular generation of PLP (Hines & Cowan, 1970).However, the interpretation of these data is made difficult by the complexity of the haematological abnormalities observed in alcoholic patients. The anaemia of alcoholism is often complicated by the simultaneous presence of folic acid deficiency, inflammation, active liver disease, and major dietary deprivations. The influence of Correspondence:Dr L. R. Solomon, Department of Medicine, University of Washington School of Medicine, Seattle, Washington 98 195. U.S.A. 0007-~048/79/030c+0356$02.00
01979 Blackwell Scientific Publications 3 43
344
L. R. Solomon and R. S . Hillman
each of these factors on vitamin B6 metabolism has yet to be determined. In addition, Chern & Beutler (1975a, b, 1976) and this laboratory (Solomon & Hillman, 1978) have reported both marked racial differences in erythrocyte PnK activity and decreases in PnK activity during red cell ageing, factors which were not recognized during previous studies. In view of these findings, an extensive study of factors affecting vitamin B6 metabolism in normal subjects and patients with iron deficiency, inflammation and alcohol associated anaemias was carried out. Red cell pyridoxine kinase (PnK) activity was determined using a new one-step microassay (Solomon & Hillman, 1976),while intracellular PLP availability was indirectly assessed by measuring erythrocyte aspartate aminotransferase (EGOT) activity (Raica & Sauberlich, 1964; Hamfelt, 1967; Sauberlich et al, 1972).The activity ofPnK and the level of intracellular PLP were found to be significantly affected by race, red cell age, pyridoxine ingestion, iron deficiency and the presence of an inflammatory state, factors which are of major importance in the interpretation of any defect of vitamin B6 metabolism in alcoholics. When these effects were taken into consideration, alcohol ingestion per se did not appear to impair either PnK activity or PLP availability. METHODS
Haemafologicat Investigations The following routine haematological tests were done: complete blood count and red cell indices by Coulter Counter, reticulocyte count, platelet count, bone marrow aspiration stained with Wright's and Prussian blue iron stains. Serum iron and total iron-binding capacity were measured by the method of Cook (1970); serum ferritin by the method of Miles et a1 (1974); serum folate by the aseptic L. casei assay of Herbert (1966);and free red cell protoporphyrin by the simplified method of Heller et a1 (1971). Red Cell Enzyme Assays Methods of cell separation, lysate preparation and PnK assay have previously been described (Solomon & Hillman, 1976; and in preparation). One unit of PnK activity is that amount of enzyme which will produce I nmole/h of pyridoxine 5-phosphate at 37°C when pyridoxine is the substrate. Since direct measurement of PLP is limited by the difficulty with quantitive extraction of PLP (Srivastava & Beutler, 1973) EGOT activity was used as an indirect measure of intracellular PLP availability. EGOT was measured by the method of Karmen et a1 (1955). One unit of EGOT activity is defined as that amount of enzyme which will produce a fall of one O.D. unit at 340 nm per minute at 37"C, using aspartic acid as substrate. Enzyme activity determined in the absence of exogenous PLP is a measure of the amount of EGOT apoenzyme saturated with PLP in vivo (saturated EGOT) while the enzyme activity determined in the presence of 0.5 rn PLP is a measure of the total active enzyme protein present in the cell (total EGOT). The degree of saturation of apoenzyme with PLP was expressed as % saturation (saturated EGOT/total EGOT x 100). Finally, the ratio of the concentrations of unsaturated EGOT/saturated EGOT, where unsaturated EGOT is the difference between total and saturated EGOT, has been shown to be a useful index of vitamin B6 nutritional status. High ratio values are associated with vitamin B6 deficiency states (Sauberlich et al, 1972). Concurrent studies on normal subjects in this laboratory indicated that both PnK and
Vitamin B6 Metabolism
345
saturated EGOT activities decrease with cell ageing (Solomon & Hillman, 1978). Significant correlations between PnK activity and both the reticulocyte count (PnK = 16.0 reticulocytes +90; r=o.78; P 0.001)were established for normal Caucasian erythrocytes fractionated into young and old populations by density centrifugation. Similarly, a significant inverse correlation was demonstrated between PnK activity and the ratio of unsaturated EGOT/saturated EGOT (PnK= -83.2 unsaturated EGOT/saturated E G O T + ~ I I r=0.71; ; P>o.ooI). Significant correlations between these same parameters were also noted in normal Blacks. The statistical significance of differences in enzyme levels in population studies were tested by means of the two-sided t-distribution. Regression analyses were performed as described by Snedecor & Cochran (1967), and differences in regression populations were analysed as described by Acton (1959). Subjects Studied Sixty-eight normals and 57 patients under investigation for haematological abnormalities were studied. Iron deficiency was considered to be present whenever the transferrin saturation fell below 15% with a serum ferritin level of less than 12 ,ug/l and absent marrow iron stores. Inflammation was identified when the patient exhibited clinical signs and symptoms of an infectious, neoplastic, or rheumatic type disorder for a minimum of 3 weeks. A pure inflammatory anaemia (anaemia of chronic disease) was diagnosed when clinical signs and symptoms of inflammation or neoplasia were present in association with a transferrin saturation greater than IS%, a reduction in total iron binding capacity, a serum ferritin greater than 12 ,ug/l, the presence of marrow iron stores and the absence of other definable causes of anaemia. Iron deficiency anaemia plus inflammation was diagnosed when the above criteria for inflammation or neoplasia were satisfied but marrow iron stores were absent. Fourteen patients ( I I Caucasians and three Blacks) were classified as having iron deficiency anaemia with or without inflammation. In seven of the Caucasian patients iron deficiency was associated with an inflammatory process. Iron loss leading to iron deficiency was attributed to peptic ulcer disease in six patients, excessive menstrual blood loss in two patients, and gastritis, eosophageal stricture, ulcerative colitis and carcinoma of the colon in one patient each. No cause was identified in the remaining two patients. The inflammatory process present in the seven Caucasian patients with iron deficiency was due to post-operative abscesses in the hip joint or pelvis (two), adenocarcinoma of the colon (one), rheumatoid arthritis (one), typhoid fever (one), toxoplasma pericarditis (one) and ulcerative colitis (one). The duration ofillness in this population ranged from I month to 4 years (mean=8 months). O f the 17 patients with an anaemia of inflammation, 12 had malignancies of various types including lymphomas (three), adenocarcinoma (three), uterine carcinoma (two), neuroblastoma (one), hypernephroma (one), ovarian carcinoma (one) and carcinoma of colon (one). The other five patients suffered from either systemic lupus erythematosus, nocardia pneumonia, fever of unknown aetiology, ischaemic colitis or Crohn’s disease. Duration of illness ranged from 2 months to 3 years (mean= 10months).
RESULTS Subjects with Iron Defects Fourteen patients with iron deficiency anaemia and 17 with the anaemia of inflammation
3 46
L. R . Solomon and R . S . Hillman
were studied as examples of disturbed iron delivery to the marrow. Routine haematologic studies and PnK and EGOT measurements are summarized in Table I. Iron deficient patients had significant elevations of PnK, total and saturated EGOT activities (Table I). Similarly, in five out of 11 Caucasians with iron deficiency the % saturation of EGOT with PLP was increased and the ratio of unsaturated EGOT/saturated was depressed, suggesting high PLP levels (Fig I ) . PnK activity was also significantly increased when expressed as units/ro9 red cells in consideration of the microcytosis and mild hypochromia of iron deficiency (PnK of 5 .o 3.4 in Caucasian iron deficient subjects and 3 . 1 & 1.2 in normals, P
Vitamin B6 Metabolism in Anaemic and Alcoholic Man L. R. SOLOMON A N D R. S. HILLMAN Department of Medicine, University o f Washington School of Medicine, Seattle, Washington and West Haven Veterans Administration Hospital and Yale University School $Medicine, New Haven, Connecticut (Received 9 November 1977; accepted for publication 28 February 1978) SUMMARY. Physiological and pathological factors affecting intracellular red cell vitamin B6 metabolism in normal, anaemic and alcoholic man were studied using a new assay for pyridoxine kinase (PnK) together with saturated and total aspartate aminotransferase (EGOT) activities as indirect indices of intracellular pyridoxal 5-phosphate (PLP) availability. In studies of anaemic states, subjects with iron deficiency anaemia demonstrated elevated levels of both PnK and saturated EGOT, while seven out of 17 subjects with inflammatory anaemia had subnormal PnK but variable saturated EGOT activities. Despite a high incidence of complicating inflammatory disease, alcoholic subjects with or without ring sideroblastic anaemia had elevated levels of both PnK and saturated EGOT. As judged from the saturated EGOT and the ratio unsaturated EGOT/saturated EGOT, intracellular PLP availability was always appropriate to the higher levels of PnK activity. Hereditary factors, inflammation, neoplasia, exposure to heavy metals, ingestion of drugs and alcoholism have all been implicated in the development of ring sideroblastic anaemias. In alcoholic patients, recent reports have suggested an association with one or more abnormalities in vitamin B6 metabolism, including: low serum levels of the active coenzyme form, pyridoxal 5-phosphate (PLP) (Leevy et al, 1965;Hines& Grasso, 1970; Eichner & Hillman, 1971;Davis & Smith, 1974; Hines & Cowan, 1974); impaired generation of plasma PLP following intravenous administration of pyridoxine (Hines & Cowan, 1970, 1974; Hines, 1975); depressed activity of erythrocyte pyridoxine kinase (PnK), the enzyme which converts the three major dietary forms of vitamin B6 to the active phosphorylated vitamins (Hines, 1975); the appearance of a plasma factor which inhibits PnK activity in vitro (Hines, 1975);and an apparent increase in a specific neutral phosphatase on the red cell membrane capable of dephosphorylating PLP (Lumeng & Li, 1974). The observation that some subjects with alcohol-induced sideroblastic anaemia fail to respond to treatment with pyridoxine but have reticulocyte responses during treatment with PLP also suggests the presence of a block in the intracellular generation of PLP (Hines & Cowan, 1970).However, the interpretation of these data is made difficult by the complexity of the haematological abnormalities observed in alcoholic patients. The anaemia of alcoholism is often complicated by the simultaneous presence of folic acid deficiency, inflammation, active liver disease, and major dietary deprivations. The influence of Correspondence:Dr L. R. Solomon, Department of Medicine, University of Washington School of Medicine, Seattle, Washington 98 195. U.S.A. 0007-~048/79/030c+0356$02.00
01979 Blackwell Scientific Publications 3 43
344
L. R. Solomon and R. S . Hillman
each of these factors on vitamin B6 metabolism has yet to be determined. In addition, Chern & Beutler (1975a, b, 1976) and this laboratory (Solomon & Hillman, 1978) have reported both marked racial differences in erythrocyte PnK activity and decreases in PnK activity during red cell ageing, factors which were not recognized during previous studies. In view of these findings, an extensive study of factors affecting vitamin B6 metabolism in normal subjects and patients with iron deficiency, inflammation and alcohol associated anaemias was carried out. Red cell pyridoxine kinase (PnK) activity was determined using a new one-step microassay (Solomon & Hillman, 1976),while intracellular PLP availability was indirectly assessed by measuring erythrocyte aspartate aminotransferase (EGOT) activity (Raica & Sauberlich, 1964; Hamfelt, 1967; Sauberlich et al, 1972).The activity ofPnK and the level of intracellular PLP were found to be significantly affected by race, red cell age, pyridoxine ingestion, iron deficiency and the presence of an inflammatory state, factors which are of major importance in the interpretation of any defect of vitamin B6 metabolism in alcoholics. When these effects were taken into consideration, alcohol ingestion per se did not appear to impair either PnK activity or PLP availability. METHODS
Haemafologicat Investigations The following routine haematological tests were done: complete blood count and red cell indices by Coulter Counter, reticulocyte count, platelet count, bone marrow aspiration stained with Wright's and Prussian blue iron stains. Serum iron and total iron-binding capacity were measured by the method of Cook (1970); serum ferritin by the method of Miles et a1 (1974); serum folate by the aseptic L. casei assay of Herbert (1966);and free red cell protoporphyrin by the simplified method of Heller et a1 (1971). Red Cell Enzyme Assays Methods of cell separation, lysate preparation and PnK assay have previously been described (Solomon & Hillman, 1976; and in preparation). One unit of PnK activity is that amount of enzyme which will produce I nmole/h of pyridoxine 5-phosphate at 37°C when pyridoxine is the substrate. Since direct measurement of PLP is limited by the difficulty with quantitive extraction of PLP (Srivastava & Beutler, 1973) EGOT activity was used as an indirect measure of intracellular PLP availability. EGOT was measured by the method of Karmen et a1 (1955). One unit of EGOT activity is defined as that amount of enzyme which will produce a fall of one O.D. unit at 340 nm per minute at 37"C, using aspartic acid as substrate. Enzyme activity determined in the absence of exogenous PLP is a measure of the amount of EGOT apoenzyme saturated with PLP in vivo (saturated EGOT) while the enzyme activity determined in the presence of 0.5 rn PLP is a measure of the total active enzyme protein present in the cell (total EGOT). The degree of saturation of apoenzyme with PLP was expressed as % saturation (saturated EGOT/total EGOT x 100). Finally, the ratio of the concentrations of unsaturated EGOT/saturated EGOT, where unsaturated EGOT is the difference between total and saturated EGOT, has been shown to be a useful index of vitamin B6 nutritional status. High ratio values are associated with vitamin B6 deficiency states (Sauberlich et al, 1972). Concurrent studies on normal subjects in this laboratory indicated that both PnK and
Vitamin B6 Metabolism
345
saturated EGOT activities decrease with cell ageing (Solomon & Hillman, 1978). Significant correlations between PnK activity and both the reticulocyte count (PnK = 16.0 reticulocytes +90; r=o.78; P 0.001)were established for normal Caucasian erythrocytes fractionated into young and old populations by density centrifugation. Similarly, a significant inverse correlation was demonstrated between PnK activity and the ratio of unsaturated EGOT/saturated EGOT (PnK= -83.2 unsaturated EGOT/saturated E G O T + ~ I I r=0.71; ; P>o.ooI). Significant correlations between these same parameters were also noted in normal Blacks. The statistical significance of differences in enzyme levels in population studies were tested by means of the two-sided t-distribution. Regression analyses were performed as described by Snedecor & Cochran (1967), and differences in regression populations were analysed as described by Acton (1959). Subjects Studied Sixty-eight normals and 57 patients under investigation for haematological abnormalities were studied. Iron deficiency was considered to be present whenever the transferrin saturation fell below 15% with a serum ferritin level of less than 12 ,ug/l and absent marrow iron stores. Inflammation was identified when the patient exhibited clinical signs and symptoms of an infectious, neoplastic, or rheumatic type disorder for a minimum of 3 weeks. A pure inflammatory anaemia (anaemia of chronic disease) was diagnosed when clinical signs and symptoms of inflammation or neoplasia were present in association with a transferrin saturation greater than IS%, a reduction in total iron binding capacity, a serum ferritin greater than 12 ,ug/l, the presence of marrow iron stores and the absence of other definable causes of anaemia. Iron deficiency anaemia plus inflammation was diagnosed when the above criteria for inflammation or neoplasia were satisfied but marrow iron stores were absent. Fourteen patients ( I I Caucasians and three Blacks) were classified as having iron deficiency anaemia with or without inflammation. In seven of the Caucasian patients iron deficiency was associated with an inflammatory process. Iron loss leading to iron deficiency was attributed to peptic ulcer disease in six patients, excessive menstrual blood loss in two patients, and gastritis, eosophageal stricture, ulcerative colitis and carcinoma of the colon in one patient each. No cause was identified in the remaining two patients. The inflammatory process present in the seven Caucasian patients with iron deficiency was due to post-operative abscesses in the hip joint or pelvis (two), adenocarcinoma of the colon (one), rheumatoid arthritis (one), typhoid fever (one), toxoplasma pericarditis (one) and ulcerative colitis (one). The duration ofillness in this population ranged from I month to 4 years (mean=8 months). O f the 17 patients with an anaemia of inflammation, 12 had malignancies of various types including lymphomas (three), adenocarcinoma (three), uterine carcinoma (two), neuroblastoma (one), hypernephroma (one), ovarian carcinoma (one) and carcinoma of colon (one). The other five patients suffered from either systemic lupus erythematosus, nocardia pneumonia, fever of unknown aetiology, ischaemic colitis or Crohn’s disease. Duration of illness ranged from 2 months to 3 years (mean= 10months).
RESULTS Subjects with Iron Defects Fourteen patients with iron deficiency anaemia and 17 with the anaemia of inflammation
3 46
L. R . Solomon and R . S . Hillman
were studied as examples of disturbed iron delivery to the marrow. Routine haematologic studies and PnK and EGOT measurements are summarized in Table I. Iron deficient patients had significant elevations of PnK, total and saturated EGOT activities (Table I). Similarly, in five out of 11 Caucasians with iron deficiency the % saturation of EGOT with PLP was increased and the ratio of unsaturated EGOT/saturated was depressed, suggesting high PLP levels (Fig I ) . PnK activity was also significantly increased when expressed as units/ro9 red cells in consideration of the microcytosis and mild hypochromia of iron deficiency (PnK of 5 .o 3.4 in Caucasian iron deficient subjects and 3 . 1 & 1.2 in normals, P
