Diversity in Intrinsic Strengths of the Human : Serum C4 Concentrations Correlate with C4 Size and Polygenic Variations, Hemolytic This information is current as Activities, and Body Mass Index of September 23, 2021. Yan Yang, Erwin K. Chung, Bi Zhou, Carol A. Blanchong, C. Yung Yu, George Füst, Margit Kovács, Ágnes Vatay, Csaba Szalai, István Karádi and Lilian Varga

J Immunol 2003; 171:2734-2745; ; Downloaded from doi: 10.4049/jimmunol.171.5.2734 http://www.jimmunol.org/content/171/5/2734 http://www.jimmunol.org/ References This article cites 65 articles, 14 of which you can access for free at: http://www.jimmunol.org/content/171/5/2734.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

by guest on September 23, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2003 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Diversity in Intrinsic Strengths of the Human Complement System: Serum C4 Protein Concentrations Correlate with C4 Gene Size and Polygenic Variations, Hemolytic Activities, and Body Mass Index1

Yan Yang,*†‡ Erwin K. Chung,*†‡ Bi Zhou,* Carol A. Blanchong,*‡ C. Yung Yu,2*†‡ George Fu¨st,2§ Margit Kova´cs,§ A´ gnes Vatay,§ Csaba Szalai,¶ Istva´n Kara´di,§ and Lilian Varga§

Among the and of the human immune system, complement component C4 is extraordinary in its frequent germline variation in the size and number of genes. Definitive genotypic and phenotypic analyses were performed on a central European Downloaded from population to determine the C4 polygenic and gene size variations and their relationships with serum C4A and C4B protein concentrations and hemolytic activities. In a study population of 128 healthy subjects, the number of C4 genes present in a diploid genome varied between two to five, and 77.4% of the C4 genes belonged to the long form that contains the endogenous retrovirus HERV-K(C4). Intriguingly, higher C4 serum protein levels and higher C4 hemolytic activities were often detected in subjects with short C4 genes than those with long genes only, suggesting a negative epistatic effect of HERV-K(C4) on the expression of C4 proteins. Also, the body mass index appeared to affect the C4 serum levels, particularly in the individuals with medium or high http://www.jimmunol.org/ C4 gene dosages, a phenomenon that was dissimilar in several aspects from the established correlation between body mass index and serum C3. As expected, there were strong, positive correlations between total C4 gene dosage and serum C4 protein con- centrations, and between serum C4 protein concentrations and C4 hemolytic activities. There were also good correlations between the number of long genes with serum levels of C4A, and the number of short genes with serum levels of C4B. Thus, the polygenic and gene size variations of C4A and C4B contribute to the quantitative traits of C4 with a wide range of serum protein levels and hemolytic activities, and consequently the power of the innate defense system. The Journal of Immunology, 2003, 171: 2734–2745.

iversities in immune functions are mainly achieved by nomenon has been observed over a decade ago (5–14), the molec- by guest on September 23, 2021 nucleotide polymorphisms/mutations and somatic re- ular genetic basis for the qualitative and quantitative variations of D combinations of gene segments (1, 2). Human comple- human complement C4 was often accounted for inaccurately. Until ment component C4 represents a new paradigm of complex, inborn recently, the sophistication of human C4 genetics has not been immune diversity (3). Phenotypically, there are acidic C4A and fully appreciated. A two-locus, C4A-C4B model with null alleles basic C4B proteins with multiple polymorphic variants in each of either locus was used to explain the great variation in the levels class (4). The plasma or serum protein levels of total C4 vary of C4A and C4B proteins in a population (15, 16). There are ac- between 100 and 1000 mg/L among different individuals, as do the tually a frequent, dichotomous gene size variation and polygenic relative quantities of C4A and C4B proteins. Although this phe- and modular duplications of C4A and C4B together with their flanking genes RP1 or RP2, CYP21A or CYP21B, and TNXA or TNXB in the MHC (Fig. 1A). Existing at the germline levels and *Center for Molecular and Human Genetics, Columbus Children’s Research Institute, †Department of Molecular Virology, Immunology, and Medical Genetics, and ‡De- inherent to different individuals, there are monomodular, bimodu- partment of Pediatrics, Ohio State University, Columbus, OH 43205; §Third Depart- lar, trimodular, or quadrimodular structures of discrete segments in ment of Internal Medicine, Semmelweis Medical University, Budapest, Hungary; and ¶Section of Molecular Immunology, Hungarian Academy of Sciences, and the MHC containing one, two, three, or four RP-C4-CYP21-TNX Heim Pa´l Pediatric Hospital, Budapest, Hungary (RCCX)3 modules (17–19). Each C4 gene may code for C4A or Received for publication February 25, 2003. Accepted for publication July 2, 2003. C4B. Each C4A or C4B gene may be 21 kb (long) or 14.6 kb The costs of publication of this article were defrayed in part by the payment of page (short) in size (20). The long C4 gene contains a 6.36-kb endog- charges. This article must therefore be hereby marked advertisement in accordance enous retrovirus HERV-K(C4) in its intron 9 (21, 22). Recently, with 18 U.S.C. Section 1734 solely to indicate this fact. we showed that the bimodular RCCX constitute 1 This work in the United States was supported by the National Institute of Arthritis slightly more than two-thirds of the Caucasian population in the and Musculoskeletal and Skin Diseases Grants R01 43969 and 1R01 AR050078, the National Institute of Diabetes and Digestive and Kidney Diseases Grant P01 midwestern United States, while the frequencies of monomodular DK55546, an institutional grant from the Columbus Children’s Research Institute and trimodular RCCX haplotypes were 0.17 and 0.14, respectively (297401), and Pittsburgh Supercomputing Center through National Institutes of (18). Quadrimodular RCCX haplotypes with four C4 genes in a Health Center for Research Resources Cooperative Agreement Grant 1P41 RR06009 (to C.Y.Y.); and in Hungary this work was supported by research grants from the Hungarian Academy of Sciences (2000-107 3,2) and Hungarian National Research Fund (T032661) (to G.F.). 3 Abbreviations used in this paper: RCCX, four-gene module in human MHC with 2 Address correspondence and reprint requests to Dr. C. Yung Yu, Children’s Re- serine/threonine nuclear protein kinase RP, complement C4, steroid 21-hydroxylase search Institute, Department of Pediatrics, The Ohio State University, 700 Children’s CYP21, and extracellular matrix protein tenascin TNX; BMI, body mass index; C4HA, Drive, Columbus, OH 43205. E-mail address: [email protected]; or Dr. George Fu¨st, C4 hemolytic activity; HERV-K(C4), human endogenous retrovirus in long C4 gene; Third Department of Medicine, Semmelweis Medical University, Budapest, HVAGE, high voltage agarose gel electrophoresis; L, RCCX module with long C4 Ku´tvo¨lgyi u´t 4, H-1125, Hungary. E-mail address: [email protected]. gene; S, RCCX module with short C4 gene.

Copyright © 2003 by The American Association of Immunologists, Inc. 0022-1767/03/$02.00 The Journal of Immunology 2735 row are rare in Caucasians, but relatively common in Asians (19, Preparation of genomic DNA, Southern blot analysis, and 23). The impacts of the C4 polygenic and gene size variations on DNA probes C4 protein levels and functional activities have not been accurately Genomic DNAs were prepared from PBMC (40). Southern blot analyses of studied in a population. TaqI-digested or PshAI-digested genomic DNAs were performed, as de- Complement component C4 is important in immunity, toler- scribed in previous publications (18, 41). The TaqI genomic blots were ance, and autoimmunity (24–27). It plays essential roles in linking hybridized with three specific probes corresponding to RP, CYP21, and the recognition pathways of the complement system initiated by TNX. The results revealed the presence and dosage of RP1-C4L (7.0 kb), RP1-C4S (6.4 kb), RP2-C4L (6.0 kb), and RP2-C4S (5.4 kb); the presence Ag-Ab or Ag-mannan-binding lectin complexes to the effectors of and relative dosage of CYP21B (3.7 kb) and CYP21A (3.2 kb); and the the humoral immune response. Complete deficiency of comple- presence and relative dosage of TNXB (2.5 kb) and TNXA (2.4 kb). The ment C4 (i.e., both C4A and C4B) is one of the most penetrant number of RCCX modules was independently confirmed by PshAI-RFLP genetic risk factors in autoimmune disease such as systemic using a RP 3Ј probe to determine the relative dosage of RP1 (7.2-kb) and RP2 (8.3-kb) fragments. The relative dosages of C4A and C4B genes were erythematosus (28, 29). Complete or partial deficiencies of C4B determined by PshAI RFLP using a C4d-specific probe corresponding to are related to vulnerability and severity of microbial infections (30, C4 gene exons 28–31 (Fig. 1). 31). Complete deficiencies of C4 in humans or guinea pigs have Determination of C4 protein polymorphism by immunofixation impairments in the secondary immune response and switching of and immunoblot analysis IgM to IgG (32–34). In an immunization process, it has been shown that low concentrations of C4 in mice (35) or partial defi- EDTA-plasma samples were digested overnight with neuraminidase at ciencies of C4A in humans are associated with higher frequencies 4°C, followed by carboxypeptidase B digest for 30 min at room tempera- ture, and resolved with high voltage agarose gel electrophoresis (HVAGE). Downloaded from of nonresponders to hepatitis B Ags (36, 37). Therefore, an accu- C4 proteins were immunofixed with goat anti-human C4 sera. Gels were rate account of the phenotypic and genotypic diversities of C4A blotted to remove diffusible proteins, and the immune complexes with C4 and C4B in a human population both qualitatively and quantita- proteins were stained with Easy-Stain (Invitrogen, Carlsbad CA), as de- tively is critically important for epidemiological studies of disease scribed previously (41–43). The relative band intensities of C4A and C4B associations with C4. allotypes in each sample were quantified by densitometry, using an Epson Expression 1600 Scanner, and analyzed by ImageQuant Version 5.0 soft- In this study, we report a comprehensive investigation of human ware. Ambiguous C4A or C4B allotypes were further investigated with im-

C4A and C4B serum protein levels and functional hemolytic ac- munoblot analysis of HVAGE-resolved gels using anti-Rg1 and anti-Ch1 http://www.jimmunol.org/ tivities in a central European population with precisely defined monoclonals provided by the VIIth Complement Genetics Workshop (4). status for C4A and C4B gene dosage, gene size, and RCCX mod- Statistical analysis ules. The study reveals higher C4 serum protein levels and hemo- lytic activities among individuals with short C4 genes than those Parametric tests (two-sample t test, paired t test, Pearson correlation anal- ysis) were used for the normally distributed (C3 and C4 levels, C4HA, with long C4 genes only. It also demonstrates that human serum BMI) data. All tests were two tailed. Multiple linear regression was used to C4 protein levels were determined by the dosage and size of C4 genes evaluate potential confounders. Statistical analysis was performed by and, unexpectedly, the body mass index (BMI) of an individual. GraphPad Prism 3.0 (GraphPad Software, San Diego, CA, www.grphpad. com) and SPSS 10.0 (SPSS, Chicago, IL) software.

Materials and Methods by guest on September 23, 2021 Results Study population Genotypic variation of human complement components C4A The study was performed in 128 healthy Hungarian subjects (40 males, 88 and C4B females). The subjects were 44.7 Ϯ 10 (mean Ϯ SD) years old. Their BMI were 25.5 Ϯ 4.7 kg/m2, and serum total cholesterol and triglyceride con- The study population consisted of 128 healthy individuals (40 centrations were 5.48 Ϯ 1.02 mM and 1.44 Ϯ 0.90 mM, respectively. males and 88 females) from Budapest, Hungary. Molecular genetic These individuals participated in a regular medical survey and gave their approaches were applied to determine the genotypic diversities of informed consent for the use of their serum/plasma and DNA samples for the constituents of RCCX modules. The total number of C4 genes the present study. For ethical reasons, after their computer registration, the data were unlinked from the subjects so that their identities could not be present in a diploid genome (i.e., gene dosage) was elucidated by traced. The study was approved by the Ethical Committee of the Semmel- TaqI RFLP that also yielded information on the details of the weis University (Budapest, Hungary). Peripheral blood samples were ob- RCCX variants such as the presence and relative dosage of RP1 tained by venipuncture and processed immediately. EDTA-plasma and se- and RP2 linked to a long or a short C4 gene, and presence and rum samples were aliquoted and stored at Ϫ80oC and thawed only immediately before the assays/measurements were performed. Genomic relative dosage of CYP21A and CYP21B, and TNXA and TNXB DNA samples were stored at 4oC. (Fig. 1B). The TaqI RFLP results were independently confirmed by PshAI RFLP analysis for the relative dosage of RP1 and RP2. Measurement of serum C4 and C3 concentrations The PshAI Southern blots were also used to determine the relative Concentrations of C4 and C3 were determined immunochemically by sin- dosage of C4A and C4B using a specific C4d genomic probe, as gle radial immunodiffusion (38), using commercial Abs against C3 and C4 one of the five nucleotide changes specific for C4A and C4B cre- (DiaSorin, Stillwater, MN). Human Serum Protein Calibrator (DAKO A/S, ated a new PshAI cleavage site in C4A (44, 45). The polymor- Glostrup, Denmark) was used as standard. phism of C4A and C4B proteins were elucidated by HVAGE of Measurement of the complement C4HA EDTA-plasma, followed by immunofixations. Selected plasma sam- C4HA were measured by the effective molecule titration method (39). ples with ambiguous C4A or C4B allotypes were further analyzed by Briefly, different dilutions of the serum samples were incubated with immunoblot analysis using anti-Rg1 and anti-Ch1 monoclonals. EAC1 at 30oC for 20 min; purified guinea pig C2 was added and further gp Demonstration of variations in the RCCX constituents. Fig. 2 incubated at 30oC for 15 min. Finally, C-EDTA (guinea pig serum diluted in buffer containing 10 mM EDTA) was added and incubated at 37oC for illustrates the genotypic characterization of the RCCX constituents 60 min. After centrifugation, OD values were read at 412 nm. C4HA in and phenotypic polymorphism of C4 in 10 individuals. Eight sub- serum was expressed in complement hemolytic units CH63 U/ml, which is jects had bimodular RCCX structures from both haplotypes; the amount of complement component required to lyse 63% of the indi- among them were homozygous LL/LL (L, long) (MK1, 11, and cator cells in standard C4 assays. Normal values of the C4HA in the lab- 13), heterozygous LL/LS (S, short) (MK2, 3, and 5), and homozy- oratory were between 5,000 and 70,000 CH63 U/ml. Hemolytic titration of C4 was standardized using aliquots stored at Ϫ70oC of the same normal gous LS/LS (MK12). Of these eight individuals who each had a human serum pool at each measurement. total of four C4 genes in a diploid genome, five had two C4A and 2736 DIFFERENT INTRINSIC STRENGTHS OF COMPLEMENT SYSTEM Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021

FIGURE 1. The gene size and polygenic variations of human complement C4 and RP-C4-CYP21-TNX (RCCX modules) in the MHC class III region. A, A molecular map of the genes in the MHC complement gene cluster. Horizontal arrows represent gene configurations; filled, vertical arrows represent locations to where the DNA probes hybridized in a TaqI genomic Southern blot analysis to determine the RCCX modular variations; gray, vertical arrows represent locations to where the DNA probes hybridized in a PshAI genomic Southern blot analysis to determine the ratio of C4A and C4B genes. B, Common RCCX length variants in healthy Caucasians. Characteristic TaqI restriction fragments of each RCCX in a Southern blot analysis are shown on the right. two C4B genes (MK2, 3, 8, 12, and 13; Fig. 2C). Two subjects, C4A6 was present in MK3. C4B1 was detectable in all except MK1 and MK5, had three C4A and one C4B. One subject (MK11) MK5, who had C4B2, and MK11, who was deficient in C4B. The had four C4A, but no C4B. The remaining two subjects had het- protein band intensities of C4A and C4B were consistent with the erozygous bimodular and monomodular RCCX structures, i.e., C4A and C4B gene dosage results. For examples, MK1 had three LS/S in MK6 and LS/L in MK9. Both of them had one C4A and C4A and one C4B genes (Fig. 2C), and the protein band intensities two C4B genes. of C4A3 allotype were about 3 times greater than that of C4B1 Results of the immunofixation experiment for C4A and C4B (Fig. 2D); MK11 had complete absence of the C4B gene (Fig. 2C) protein allotypes are shown in Fig. 2D. In addition to C4A3 that and that was matched with the absence of C4B protein (Fig. 2D). was observed in every individual, C4A2 was present in MK2 and The TaqI and PshAI RFLP patterns for the RCCX modules The Journal of Immunology 2737

FIGURE 2. Genotypic variations of RCCX modules, C4A and C4B gene dosages, and protein polymorphisms. A, TaqI RFLP to show RCCX modular variations. Hybridiza- tion probes include 3Ј RP cDNA, CYP21 genomic, and 3Ј TNX genomic fragments. B, PshAI-RFLP to show number of RCCX mod- ules by determining the relative dosages of RP1 and RP2.A3Ј RP cDNA probe was used for hybridization. C, PshAI-RFLP to show relative dosages of C4A and C4B genes. Hy- Downloaded from bridization probe used was a genomic frag- ment corresponding to C4 exons 28–31. D, Allotyping of human C4A and C4B proteins. EDTA plasma C4 proteins were resolved by HVAGE, processed by immunofixation using goat polyclonal antisera against human C4.

Notice the presence of restriction fragments http://www.jimmunol.org/ corresponding to CYP21B and TNXB only (marked by arrows, A), and the presence of a novel 3.5-kb PshAI fragment corresponding to a rearranged TNXA-XB recombinant (B)in subject MK6. Subject MK11 was completely C4B deficient, as there were no C4B gene (marked by an arrow, C) and no C4B protein (marked by an arrow, D). by guest on September 23, 2021

revealed a common pattern for bimodular structures with RP1-C4- of the trimodular haplotypes LLL, LSS, and LSL were 0.039, CYP21A-TNXA-RP2-C4-CYP21B-TNXB, and a common pattern 0.027, and 0.004, respectively. Altogether, the frequencies of for monomodular structures with RP1-C4-CYP21B-TNXB (Fig. monomodular, bimodular, and trimodular RCCX haplotypes were 1B). A single anomaly was observed in MK6, whose markers for 0.133, 0.797, and 0.070, respectively. CYP21A and TNXA in the TaqI RFLP appeared to be absent, al- In this central European population, the number of C4 genes in though he had a bimodular LS and a monomodular S structure. The a diploid genome varied from two to five. The majority of indi- PshAI RFLP hybridized to RP 3Ј probe revealed the presence of a viduals had four C4 genes per genome, which had a frequency of 3.5-kb, rearranged fragment that is characteristic of RP2-TNXA ϩ 0.656. Individuals with C4 gene dosages of three and five had 120. In other words, MK6 has an unusual LS haplotype that is frequencies of 0.211 and 0.117, respectively. Only two individuals characterized by the presence of two CYP21B genes and a rear- had a dosage of two C4 genes (frequency: 0.016). We did not ranged TNXA with the 120-bp sequence that is normally present in detect any individuals in the study population with six C4 genes. exon 36-intron 36 of the TNXB gene (i.e., RP1-C4-CYP21B-TNXA In the 128 genotyped individuals, there were 496 C4 genes. ϩ 120-RP2-C4-CYP21B-TNXB) (46). Among these genes, 77.4% belonged to the long form with HERV- The RCCX and complement C4 polygenic variations in the K(C4) integrated into the ninth intron; 22.6% belonged to the short Budapest population. The frequencies of RCCX length variants, form without the endogenous retrovirus in the C4 gene. The av- namely, monomodular L and S; bimodular LL and LS; and trimo- erage C4 gene number in the normal Hungarian population (gene dular LLL, LSS, and LSL, are listed in Table I. The most common index) was 3.875, of which C4A had a gene index of 2.0 and C4B, RCCX haplotype in the Hungarian population was the bimodular 1.875. The frequency of individuals with gene dosage of two C4A LL with a frequency of 0.500. Another bimodular haplotype LS was 0.570, and with two C4B was 0.700. Overall, one-fifth of the had a frequency of 0.289. The frequencies for monomodular L and study population had a single C4A gene in a diploid genome; the S haplotypes were 0.047 and 0.086, respectively. The frequencies same frequency was also found in subjects with single C4B genes. 2738 DIFFERENT INTRINSIC STRENGTHS OF COMPLEMENT SYSTEM

Table I. A summary of RCCX length variants and C4 gene dosages in had no C4A gene (frequency: 0.016), and one individual had no Caucasians C4B gene (frequency: 0.008).

Hungary Ohioa Qualitative variations of C4A and C4B

n Frequency Frequency The polymorphism of C4A and C4B proteins was investigated by immunofixation of EDTA-plasma, resolved by HVAGE. Samples RCCX length variants S 22 0.086 0.110 showing ambiguous or unusual C4A and C4B allotypes were fur- L 12 0.047 0.060 ther analyzed by immunoblot analysis using anti-Rg1 and anti-Ch1 LL 128 0.500 0.460 mAbs. The frequencies of C4A and C4B allotypes detected in this LS 74 0.289 0.230 Hungarian study population are listed in Table II. LLL 10 0.039 0.073 Among the 256 C4A genes in the study population, 249 coded LSL 1 0.004 0.003 LSS 7 0.0273 0.063 for a detectable protein. The most common C4A allotype was A3, Lxb 2 0.008 0 which had a frequency of 85.5%. C4A2, and A6 had frequencies Total 256 1.00 1.00 of 7.81 and 1.95%, respectively. Rare C4A allotypes detected in Long (L) and short (S) C4 genes the study population included A92, A12, A4, and A5, which had a L genes 384 0.774 0.761 S genes 112 0.226 0.239 combined frequency of 1.95%. In the study population, there were Total 496 1.00 1.00 seven suspected C4A mutant genes not producing C4A proteins. Total number C4 genes in a diploid genome (C4 gene dosage) Among them, three (1.17%) were confirmed by sequence-specific 2 2 0.016 0.027 primer PCR to have nonsense mutations caused by a 2-bp insertion Downloaded from 3 27 0.211 0.273 at exon 29 (47, 48). 4 84 0.656 0.493 5 15 0.117 0.173 In the study population, all 240 C4B genes coded for C4B pro- 6 0 0 0.033 teins, of which C4B1 had a frequency of 83.8%. C4B2 and B92 Total 128 1.00 1.00 had frequencies of 10.8 and 1.25%, respectively. Rare C4B allo- Total C4 gene index 3.875 3.933 types detected include B94, B12, B15, B3, B4, B5, and B6, which

C4A gene dosage http://www.jimmunol.org/ 0 2 0.016 0.007 had a combined frequency of 4.17%. 1 26 0.203 0.200 2 73 0.570 0.460 Quantitative diversities of C4A and C4B proteins 3 24 0.188 0.295 4 3 0.023 0.033 The C4 protein level in each individual was measured by single 5 0 0 0.007 radial immunodiffusion of serum proteins. In the protein-allotyp- Total 128 1.00 1.00 ing gels, the band intensities for C4A and C4B were quantified by C4A gene index 2.0 2.17 densitometry and the corresponding C4A and C4B isotype protein C4B gene dosage levels calculated from the total serum C4 concentrations. The re- 0 1 0.008 0.020

1 26 0.203 0.320 lationship of total C4, C4A, and C4B serum concentrations with by guest on September 23, 2021 2 90 0.703 0.533 respect to the dosages of total C4 (C4A plus C4B), C4A, C4B, long 3 10 0.078 0.120 C4 genes, and short C4 genes was analyzed and shown in Fig. 3. 4 1 0.008 0.007 For total C4 gene dosage, the population was categorized into three Total 128 1.00 1.00 Ͻ C4B gene index 1.875 1.77 groups: individuals with 4 C4 genes (medium), 4 C4 genes (2 or 3; low), and Ͼ4 C4 genes (5; high). With respect to C4A gene a Taken from Ref. 18. Ͻ b Undetermined due to a Taq1 RFLP at the 5Ј end of the second C4 gene. dosage, the population was grouped into individuals with 2 C4A genes (0 or 1; low), and Ն2 C4A genes (2, 3, or 4; high). The same categorization was applied for C4B genes. With respect to long C4 The frequency of individuals with a gene dosage of three C4A was genes, the population was divided into two groups: 0–2 long C4 0.190, and that of three C4B was 0.080. Three individuals (fre- genes, and 3–5 long C4 genes. With respect to short C4 genes, the quency: 0.023) had four C4A genes, and one single individual population was divided into two groups: with (i.e., 1–3) and with- (frequency: 0.008) had four C4B genes in a genome. Two subjects out (i.e., 0) short C4 genes.

Table II. Phenotypes of complement components C4A and C4B in a healthy Hungarian population

Frequency (%) Frequency (%) Allotype Allotype C4A n C4A All C4a C4B n C4B All C4a

A2 20 7.81 4.03 B1 201 83.8 40.5 A3 219 85.5 44.2 B2 26 10.8 5.24 A4 1 0.39 0.20 B92 3 1.25 0.60 A5 2 0.78 0.40 B4 2 0.83 0.40 A6 5 1.95 1.01 B5 2 0.83 0.40 Othersb 2 0.78 0.4 Othersc 6 2.5 1.21 AQ0d 7 2.73 1.41 BQ0d 00 0 (2-bp ins)e (3) (1.17) (0.60) Total 256 100 51.6 Total 240 100 48.4

a Among all C4 (C4A plus C4B) allotypes. b A12, A92-Ch1. c B94, B12, B15, B3, B6, B7. d Mutant C4; no C4 protein produced from a C4A or C4B gene. e C4Q0 mutant caused by 2-bp insertion at codon 1213 from exon 29. The Journal of Immunology 2739 Downloaded from http://www.jimmunol.org/ by guest on September 23, 2021

FIGURE 3. Analyses of C4 serum concentrations with C4 gene dosage and gene size variations. AÐE, Show the relationships of total serum C4 levels with C4 polygenic and gene size variations. FÐJ, Show the relationships of serum C4A proteins with C4 polygenic and gene size variations. KÐO, Show the relationships of serum C4B proteins with C4 polygenic and gene size variations. A, F, and K, The p values for Tukey post hoc test of one-way ANOVA are indicated. For all other panels, p values for unpaired t test are indicated. 2740 DIFFERENT INTRINSIC STRENGTHS OF COMPLEMENT SYSTEM

Total serum C4 protein concentrations and C4 gene dosages. As C4 serum concentrations (Fig. 4F), and to the serum levels of C4A shown in Fig. 3A, there was a positive correlation of serum C4 (Fig. 4G) and C4B (Fig. 4H)(p Ͻ 0.0001). The correlation coef- concentrations with total C4 gene dosages. The low C4 gene dos- ficients were ϳ0.5. The strengths of correlation between C4HA age group (gene dosages, 2–3; n ϭ 29) had a mean C4 concen- and C4A protein levels, and between C4HA and C4B protein lev- tration of 0.33 Ϯ 0.10 g/L, which was significantly lower than that els were similar. of the medium gene dosage group (gene dosage, 4; n ϭ 84; con- To further illustrate the intricate relations among the size of C4 centration 0.46 Ϯ 0.15 g/L; p Ͻ 0.001) and the high gene dosage genes, C4 serum protein concentrations, and hemolytic activities, group (gene dosage, 5; n ϭ 15; concentration 0.50 Ϯ 0.16 g/L; p Ͻ we focused on the most common gene dosage group, i.e., subjects 0.001). The total serum C4 concentration was significantly higher with four C4 genes (two C4A and two C4B) and organized in one in individuals with 2, 3, or 4 C4A genes (concentration 0.46 Ϯ 0.16 of the following three RCCX configurations: LL/LL, LL/LS, and g/L; n ϭ 100; p ϭ 0.0196) than in those with 0 or 1 C4A gene LS/LS. In comparison with LL/LL homozygotes, individuals with (concentration 0.38 Ϯ 0.13 g/L; n ϭ 28) (Fig. 3B). With respect to LS/LS structures had higher serum C4 protein concentrations ( p Ͻ the number of the C4B genes, however, there was no relevant 0.05; Fig. 5A) and higher C4 hemolytic activities ( p Ͻ 0.01; Fig. correlation with the total C4 serum concentration (Fig. 3C). With 5B). Although not reaching a statistically significant level, het- respect to the dosage of the long C4 genes, there was also no erozygous LL/LS had values between LL/LL and LS/LS for both relevant difference in the C4 serum concentrations between the low serum C4 protein concentrations and C4HA. Compared with those dosage and the high dosage groups (Fig. 3D). On the contrary, a having homozygous LL/LL, the C4A and C4B serum protein lev- higher total C4 serum concentration was observed in the popula- els increased by 26.0 and 34.0%, respectively, in LL/LS heterozy- tion group with one or more short C4 genes (concentration 0.47 Ϯ gotes, and by 34.0 and 40.0% in LS/LS homozygotes. The increase Downloaded from 0.17 g/L; n ϭ 80) than the group without any short C4 genes of C4B protein levels was significant in both LL/LS ( p Ͻ 0.01) (concentration 0.39 Ϯ 0.11 g/L; n ϭ 48; p ϭ 0.0095; Fig. 3E). and LS/LS ( p Ͻ 0.01). This phenomenon probably suggested that Serum C4A and C4B concentrations and C4 gene dosages. We the presence of short C4 gene(s) at the second locus not only then analyzed the impact of gene dosages for total C4, C4A, C4B, significantly increased the expression of the C4B proteins, but also long genes, and short genes on the serum protein concentrations of had a considerable, positive impact on the expression of C4A pro-

C4A and C4B. The mean C4A serum concentrations in the low teins. It appeared that the presence of HERV-K(C4) reduced the http://www.jimmunol.org/ (i.e., Ͻ2), medium (i.e., ϭ 2), and high (i.e., Ͼ2) C4 gene dosage expression of serum C4 and the C4 hemolytic activities in a dose- groups were 0.13 Ϯ 0.05, 0.23 Ϯ 0.08, and 0.26 Ϯ 0.08 g/L, dependent manner. respectively. There was a close, positive correlation of C4A serum concentrations with the total C4 gene dosage ( p Ͻ 0.001; Fig. 3F). BMI and serum C4 and C3 concentrations As expected, the serum C4A concentration was a function of C4A gene dosage ( p Ͻ 0.0001; Fig. 3G), and was weakly, but inversely In view of the emerging roles of several complement proteins in related to the C4B gene dosage (Fig. 3H). With respect to C4 gene lipid metabolism and the strong correlation of serum C3 levels and size, higher serum C4A concentrations were present in individuals BMI in healthy people, we investigated whether there were posi- with high dosage (concentration 0.23 Ϯ 0.08 g/L) than those with tive relationships between BMI and C4 gene dosage; total C4, by guest on September 23, 2021 low dosage of long C4 genes (concentration 0.18 Ϯ 0.09 g/L; p ϭ C4A, C4B serum protein concentrations; and C4 HA (Fig. 6). BMI 0.0028; Fig. 3I). The presence of short C4 genes, however, did not was not related to the C4 gene dosage or C4HA (Fig. 6, A and E, appear to have a correlation with the C4A protein concentration respectively). However, there was a modest, but significantly pos- ( p ϭ 0.9392; Fig. 3J). itive correlation between BMI and serum protein concentrations of The mean C4B serum concentrations were positively correlated C4, C4A, and C4B (Fig. 6, B, C, and D, respectively). with C4B gene dosage ( p Ͻ 0.0001; Fig. 3M), and a highly sig- As expected, C3 serum levels were found to be strongly corre- nificant correlation was noted with the presence of one or more lated to the BMI (Fig. 6F). Approximately the same coefficients short C4 genes ( p Ͻ 0.0001; Fig. 3O). The serum C4B level was were found in the males and females for each correlation depicted not related to C4A gene dosage (Fig. 3L). There was a mild, but in Fig. 6. The only exception was the correlation between serum significant inverse correlation of serum C4B levels with the num- C3 concentrations and BMI. In males, we found a very strong ber of long C4 genes ( p ϭ 0.048; Fig. 3N). Unlike the case for correlation (r ϭ 0.675, p ϭ 0.0004), whereas in females the cor- C4A, there was no significant correlation of C4B protein concen- relation was weaker (r ϭ 0.346, p ϭ 0.0041). trations with total C4 gene dosages (Fig. 3K). The serum C4 concentration did not correlate to the age of sub- jects or the serum total cholesterol levels (correlation coefficient r ϭ 0.091, p ϭ 0.316) and triglyceride (r ϭ 0.082, p ϭ 0.362). By Hemolytic activities, C4 gene dosages, C4 gene sizes, and serum contrast, there was a significant positive correlation between C3 C4 concentrations levels and the total cholesterol concentration (r ϭ 0.246, p ϭ C4HA of the serum samples were determined by effective mole- 0.015), and a very strong correlation between C3 levels and con- ϭ Ͻ cule titration and expressed in CH63 U/ml. The C4HA from each centrations of triglycerides (r 0.510, p 0.001). sample was scatter plotted against the C4 gene dosages (Fig. 4). Using multiple linear regression analysis, we found the corre- There was a direct correlation of the mean C4HA with an increase lation between BMI and serum C4 concentrations to be indepen- of total C4 gene dosage. This was most obvious when the com- dent from that between BMI and serum C3 concentrations (Table parison was made between the high C4 gene dosage group and the III). In the whole population, a strong correlation between C3 and low gene dosage group ( p Ͻ 0.01; Fig. 4A), and also between the BMI was found when no adjustment was applied. However, when groups with one to three short C4 genes and without short C4 serum cholesterol and triglyceride levels were adjusted to the age, genes ( p ϭ 0.0063; Fig. 4E). However, there was no clear-cut gender, cholesterol, and triglyceride levels, the correlation coefficient correlation of C4HA with the gene dosage of C4B (Fig. 4C)or between C3 concentration and BMI became marginally significant. with the dosage of long C4 genes (Fig. 4D). This difference became more pronounced when the analysis was re- When the relations of serum C4 concentrations and C4HA were stricted to the subjects with only bimodular RCCX modules. In this examined, it was found that C4HA strongly correlated to the total group, no significant correlation ( p ϭ 0.356) was noted between C3 The Journal of Immunology 2741 Downloaded from

FIGURE 4. Analyses of C4 hemolytic activi- ties with C4 polygenic and gene size variations (AÐE), and with serum C4 concentrations (FÐH). Serum hemolytic activity of C4 expressed in http://www.jimmunol.org/ CH63 U/ml in the sera of 128 healthy subjects with low (gene dosages ϭ 2–3), medium (gene dosage ϭ 4), or high (gene dosage ϭ 5) dosages of C4 genes (A); low (gene dosage ϭ 0or1)or high (gene dosages ϭ 2–4) number of C4A genes (B); or C4B genes (C), low (gene dosages ϭ 0–2) or high (gene dosages ϭ 3 to 5) number of long (L) C4 genes (D), and none or 1–3 short (S) C4 genes (E). A, The p values for Tukey post hoc test of one-way ANOVA are indicated. BÐE, The p by guest on September 23, 2021 values for unpaired t test are indicated. FÐH, The Pearson correlation coefficients and their signifi- cance are indicated.

levels and BMI after adjustment. Similar disappearance of the corre- lipid levels did not negatively affect the strength of the correlation lation between C3 and BMI ( p ϭ 0.318) after adjustment was noted between C4 levels and BMI and in the group of subjects with at least in subjects with at least four RCCX modules (i.e., those with mono- four RCCX modules: the adjusted correlation between C4 concentra- modular/monomodular and monomodular/bimodular RCCX combi- tion and BMI became highly significant instead ( p ϭ 0.009) nations were excluded). By contrast, adjustments to age, gender, and (Table III). 2742 DIFFERENT INTRINSIC STRENGTHS OF COMPLEMENT SYSTEM

FIGURE 5. The impact of long and short C4 genes on the serum C4 protein concentrations and C4HA. Eighty healthy subjects with bimodu- lar/bimodular RCCX haplotypes cor- responding to LL/LL, LL/LS, or LS/LS configurations were chosen for analysis. The p values for Tukey post hoc test of one-way ANOVA are indicated. Downloaded from http://www.jimmunol.org/

Discussion genetics, which we suggest to be an adaptation of the innate im- In this study, we report three important novel findings. First, we mune system to provide a wide spectrum of immune efficiencies demonstrated that the serum concentrations of C4A and C4B pro- and strengths to defend against parasites. We have established the teins and total amounts of C4 proteins were determined by the sophistication of the C4 genetics in a Caucasian population in mid- number of long and short C4 genes. Second, in individuals with western region of the United States (18). This current study does medium or high C4 gene dosages (i.e., Ն4 genes), a significant not only support and validate results of the Ohioan study, it also positive correlation was found between BMI and C4 serum levels, provides an essential link between the C4 genetic diversity with which in several aspects was quite different from the correlation the phenotypic and functional variations. The C4 gene number and by guest on September 23, 2021 between BMI and serum C3 concentrations. Third, by using pre- gene size and, unexpectedly, the BMI are all relevant in determin- cise and definitive methods in genotyping and phenotyping exper- ing the C4 serum protein concentrations, which impact the power iments, we demonstrated the correlation between C4 gene dosage of the complement system as reflected by the hemolytic assays. In and C4 protein concentrations. The results provide a foundation for a separate report, we have delineated in the same Caucasian pop- the differences on an important factor of innate immune response. ulations the complex relationship between complement C4 and They enable a better understanding of the genetic basis for the RCCX modular variations with two common single nucleotide poly- qualitative and quantitative diversities of human complement C4, morphisms of the TNF gene TNFA, Ϫ238 G/A and Ϫ308 G/A (56). which include variations in immunochemical and functional prop- The serum C4A protein levels were closely associated with C4A erties of C4A and C4B, and a wide range of serum protein levels gene dosage, the number of long C4 genes, and the total C4 gene and hemolytic activities. dosages. This phenomenon may reflect the fact that most C4A genes are located in the first RCCX module that frequently con- Factors determining the serum C4 protein levels tains the endogenous retrovirus HERV-K(C4), and an increase in Under pathological conditions, low serum or plasma C4 levels may C4 gene dosage has a higher tendency to increase the number of be caused by excessive consumptions due to complement activa- C4A genes. In contrast, serum C4B protein levels were mainly the tion as a result of microbial infections or abnormalities such as a functions of gene dosages of C4B and short C4 genes. The latter deficiency of the complement C1 inhibitor (49, 50), and the pres- reflects the fact that more short C4 genes code for C4B proteins ence of nephritic factors or autoantibodies that stabilizes the C1 than for C4A proteins. complex (51, 52). Therefore, the serum concentrations of total C4 and C3 and their split products have been used as indicators for The significance and implication of the long and short C4 genes disease activities of autoimmune disease, and inflammatory re- It is of interest to note that individuals containing one or more sponses including graft rejections (53–55). However, there is a short C4 genes consistently have higher serum total C4 and C4B more fundamental and sophisticated genetic mechanism leading to concentrations than those with long C4 genes only. Higher expres- a wide range of serum protein levels of total C4, C4A, and C4B in sion levels of serum C4B proteins than C4A proteins in individuals the general population. Besides the polygenic variation from two with equal number of C4A and C4B genes had been observed to five genes among different individuals in our study population, previously in specific haplotypes such as HLA B18 DR2 C4A4 B2 the gene size dichotomy adds an additional dimension to the ge- (57). This B18 DR2 haplotype has a bimodular RCCX structure netic complexity of human C4. Close to three-quarters (77.4%) of LS, and it is plausible that the short C4B gene is responsible for the the C4 genes in this study population were long, and about one- higher serum level of the C4B2 protein than the C4A4 protein. quarter (22.6%) of the C4 genes were short. This intrinsic diversity However, the positive impact of short C4 genes seemed more far in the C4 gene number and gene size is extraordinary in human reaching because it affected not only the C4B (and the total C4 The Journal of Immunology 2743

FIGURE 6. Analyses of BMI and C4 polygenic variations (A), serum C4 concentrations (BÐD), C4HA (E), and serum C3 concentration (F). A, The p values

for Tukey post hoc test of one-way ANOVA are in- Downloaded from dicated. BÐF, The Pearson correlation coefficients and their significance are indicated. BMI was expressed in kg/m2. http://www.jimmunol.org/ by guest on September 23, 2021

protein levels), but also the levels of C4A proteins, albeit to a have serum C4 protein levels 41.0% higher than those with lesser extent. This was best illustrated in subjects with bimodular LLL/LL. RCCX structures from both copies of 6 and con- The selection advantage of having an endogenous retrovirus in tained two C4A and two C4B genes. Individuals with homozygous three-quarters of C4 genes in the Caucasian population is a matter LS/LS had mean serum C4A and C4B protein levels 34.0 and of speculation. It was proposed that the reversely oriented HERV- 40.0% higher than those from homozygous LL/LL, respectively. K(C4) in the intron 9 of the long C4 gene would lead to the pro- The levels of C4A and C4B proteins from individuals with het- duction of an antisense transcript of the ancient retrovirus when- erozygous LL/LS were higher than those with LL/LL, but lower ever the long C4 gene is transcribed (21). This antisense retroviral than those with LS/LS (Fig. 5). It is likely that the 6.36-kb endog- RNA transcript with multiple defective mutations could confer the enous retrovirus in the long C4 genes retarded the gene transcrip- host selection advantage in the defense against retroviral infec- tion rate because of the larger size of the heteronuclear transcripts, tions. This is because partial hybridizations of the antisense en- and consequently reduced the corresponding protein products. dogenous retroviral RNA transcripts to the proviral RNA genomes Consistent with this notion, we found that individuals with LSS/LL would trigger the cellular RNase surveillance and the IFN systems

Table III. Linear regression analysis of the relationship between BMI and serum C3 and C4 concentrations in healthy Hungarian subjects

In Subjects with Bimodular/Bimodular In Subjects with at Least Four RCCX In the Entire Study Populationa (n ϭ 128) RCCX Modulesa (n ϭ 80) Modules in a Diploid Genomea (n ϭ 99)

Unadjusted Adjustedb Unadjusted Adjustedb Unadjusted Adjustedb

C3 5.384 (Ͻ0.001) 2.580 (0.077) 4.978 (0.014) 1.904 (0.356) 4.398 (0.010) 1.733 (0.318) C4 5.249 (0.073) 5.049 (0.059) 8.812 (0.028) 8.066 (0.031) 8.801 (0.013) 8.645 (0.009)

a Partial regression coefficients are shown; p values are shown in parentheses. b Adjusted to age, gender, cholesterol, and triglyceride levels. 2744 DIFFERENT INTRINSIC STRENGTHS OF COMPLEMENT SYSTEM to degrade the double-stranded retroviral hybrids with multiple mis- C4A and C4B genes and their serum levels may be helpful. This is matches. Thus, the host endogenous retroviruses could be a form of because the C4 data would reflect the strength of innate immunity, innate defense system against retroviral infections (21, 22). and also the vulnerability to infections and susceptibility to auto- immune diseases. It is important to note that the serum C4A and C4HA as a measure of the strength of the complement system C4B are predominantly secreted by the liver and the results pre- Results of our experiments on the functional activities of C4A and sented in this work largely reflected how the C4 gene size and C4B using a conventional hemolytic assay revealed a direct cor- polygenic variations and the BMI influenced the production of C4 relation of the hemolytic activities with serum C4 protein concen- by the liver. The influence of C4 polygenic and gene size varia- trations. Higher C4 gene dosage and presence of short C4 gene(s) tions on the regulation mechanisms of C4A and C4B protein bio- increased the serum C4 protein levels and hemolytic reactivities. synthesis in local tissues such as the kidneys, heart, CNS, thyroids, Importantly, this observation reflects the influence of the gene size and adrenals is yet to be determined. and polygenic variations of human C4 on the intrinsic strength of the complement system. Acknowledgments In in vitro studies using purified human C4 proteins in most We sincerely thank the blood donors for participation in this study. We are heterologous assays using rabbit Abs and C4-deficient guinea pig indebted to Charlene Cameron for excellent secretarial assistance. complement proteins, activated human C4A proteins have high References binding affinities to amino group-containing Ags or immune 1. Litman, G. P., M. K. Anderson, and J. P. Rast. 1999. Evolution of antigen binding complexes; activated C4B proteins have an internal catalytic receptors . Annu. Rev. Immunol. 17:109. mechanism that favors rapid covalent binding to hydroxyl group- 2. Wagner, S. D., and M. S. Neuberger. 1996. Somatic hypermutation of immuno- Downloaded from containing Ags through a transesterification mechanism (58, 59). globulin genes. Annu. Rev. Immunol. 14:441. 3. Yu, C. Y., E. K. Chung, Y. Yang, C. A. Blanchong, N. Jacobsen, K. Saxena, Unexpectedly, the C4 hemolytic activities from 128 samples using Z. Yang, W. Miller, L. Varga, and G. Fust. 2003. Dancing with complement C4 conventional hemolytic assays were indifferent to the C4A and and the RP-C4-CYP21-TNX (RCCX) modules of the major histocompatibility complex. Prog. Nucleic Acid Res. Mol. Biol. 75:217. C4B isotype composition in the serum. In other words, serum C4A 4. Mauff, G., B. Luther, P. M. Schneider, C. Rittner, B. Strandmann-Bellinghausen, and C4B proteins manifested similar activities in the fluid-phase R. Dawkins, and J. M. Moulds. 1998. Reference typing report for complement component C4. Exp. Clin. Immunogenet. 15:249. hemolytic assays, if their serum protein levels were the same. In- http://www.jimmunol.org/ 5. O’Neill, G. J., and B. DuPont. 1979. Serum C4 levels, Chido, Rodgers, and deed, differential reactivities between C4A and C4B for bindings allotypes of C4 component of complement. Transplant. Proc. 11:1102. to immune complexes, human or sheep erythrocytes, and comple- 6. Averill, B. K., and J. E. Bernal. 1984. Genetic and environmental influences on ment receptor 1, and for prevention of immunoprecipitation and serum levels of human C4. Ann. Hum. Biol. 11:149. 7. Welch, T. R., L. Beischel, A. Berry, J. Forristal, and C. D. West. 1985. The effect solubilization of immune aggregates could be demonstrated only of null C4 alleles on complement function. Clin. Immunol. Immunopathol. 34: when the C4A and C4B proteins were purified by an immuno- 316. 8. Uko, G., F. T. Christiansen, R. L. Dawkins, and V. J. McCann. 1986. Reference chemical method or resolved by gel electrophoresis (16, 60–65). ranges for serum C4 concentrations in subjects with and without C4 null alleles. A differential reactivity between C4A and C4B could not be de- J. Clin. Pathol. 39:573. tected in assay systems when whole human sera were used as a 9. Holme, E., S. J. Cross, J. Veitch, G. J. O’Neill, and K. Whaley. 1988. Quanti-

tation of human C4A and C4B, in serum and plasma by enzyme-linked immu- by guest on September 23, 2021 source of C4A or C4B proteins or in reconstituted assays when noadsorbent assay. Immunogenetics 27:295. human C4A- or C4B-deficient sera or human sera with selectively 10. Teisberg, P., R. Jonassen, B. Mevag, T. Gedde-Dahl, Jr., and B. Olaisen. 1988. depleted C4 were used (64, 66–68). One explanation is that the Restriction fragment length polymorphisms of the complement component C4 loci on : studies with emphasis on the determination of gene num- native or activated C4A and/or C4B proteins are regulated or pro- ber. Ann. Hum. Genet. 52:77. tected by protein(s) in the autologous sera, which mask their dif- 11. Chrispeels, J., B. Stephanie, C. Rittner, and D. Bitter-Suermann. 1989. Sandwich enzyme-linked immunosorbent assays for the quantification of the C4 isotypes ferential chemical reactivities in a conventional assay system. (C4A and C4B) in human plasma. J. Immunol. Methods 125:5. 12. Moulds, J. M., F. C. Arnett, C. M. Giles, and R. G. Hamilton. 1990. A novel Correlation between BMI and serum concentrations of C4 immunoassay for the quantitation of human C4 gene products. Complement In- flamm. 7:95. The observation of a correlation between human serum C4 levels 13. Hammond, A., W. Ollier, and M. J. Walport. 1992. Effects of C4 null alleles and and BMI is novel. When calculated by multiple linear regression homoduplications on quantitative expression of C4A and C4B. Clin. Exp. Immu- analysis adjusted to sex, age, and serum lipid levels, the correlation nol. 88:163. 14. Rebmann, V., I. Doxiadis, B. S. Kubens, and H. Grosse-Wilde. 1992. Quantita- of C4 and BMI was stronger in the subjects with bimodular RCCX tion of the human component C4: definition of C4 Q0 alleles and C4A duplica- modules only than in the entire study population, and it was the tions. Vox Sang. 62:117. 15. O’Neill, G. J., S. Y. Yang, and B. DuPont. 1978. Two HLA-linked loci control- strongest when subjects with monomodular/monomodular and ling the fourth component of human complement. Proc. Natl. Acad. Sci. USA monomodular/bimodular combinations were excluded from the 75:5165. analysis (Table III). The present findings indicate that the mech- 16. Awdeh, Z. L., and C. A. Alper. 1980. Inherited structural polymorphism of the fourth component of human complement. Proc. Natl. Acad. Sci. USA 77:3576. anism of correlation between C4 and BMI is independent and dif- 17. Yang, Z., A. R. Mendoza, T. R. Welch, W. B. Zipf, and C. Y. Yu. 1999. Modular ferent from that of complement C3 and BMI. In contrast to C3, variations of HLA class III genes for serine/threonine kinase RP, complement C4, there is no correlation between serum lipid levels and serum C4 steroid 21-hydroxylase CYP21 and tenascin TNX (RCCX): a mechanism for gene deletions and disease associations. J. Biol. Chem. 274:12147. concentrations. Moreover, at multiple linear regression analysis, 18. Blanchong, C. A., B. Zhou, K. L. Rupert, E. K. Chung, K. N. Jones, J. F. Sotos, C3 and BMI did not significantly correlate after adjustment to R. M. Rennebohm, and C. Y. Yu. 2000. Deficiencies of human complement component C4A and C4B and heterozygosity in length variants of RP-C4- serum lipid levels; the same adjustment did not negatively influ- CYP21-TNX (RCCX) modules in Caucasians: the load of RCCX genetic diver- ence the correlation between C4 serum levels and BMI. The cor- sity on MHC-associated disease. J. Exp. Med. 191:2183. relation between C3 and BMI was stronger in the males than in the 19. Chung, E. K., Y. Yang, R. M. Rennebohm, M. L. Lokki, G. C. Higgins, K. N. Jones, B. Zhou, C. A. Blanchong, and C. Y. Yu. 2002. Genetic sophisti- females, whereas no gender-related difference could be detected in cation of human complement C4A and C4B and RP-C4-CYP21-TNX (RCCX) the relationship of C4 levels and BMI. Why the serum C4 levels modules in the major histocompatibility complex (MHC). Am. J. Hum. Genet. are higher in obese than in lean subjects remains to be investigated. 71:823. 20. Yu, C. Y. 1991. The complete exon-intron structure of a human complement In conclusion, we have shown the C4 polygenic and gene size component C4A gene: DNA sequences, polymorphism, and linkage to the 21- variations contributing to a wide range of serum levels for C4A hydroxylase gene. J. Immunol. 146:1057. 21. Dangel, A. W., A. R. Mendoza, B. J. Baker, C. M. Daniel, M. C. Carroll, L.-C. and C4B. This phenomenon is relevant. When a patient is diag- Wu, and C. Y. Yu. 1994. The dichotomous size variation of human complement nosed and treated for a disease, the information on the number of C4 gene is mediated by a novel family of endogenous retroviruses which also The Journal of Immunology 2745

establishes species-specific genomic patterns among Old World primates. Immu- 46. Rupert, K. L., R. M. Rennebohm, and C. Y. Yu. 1999. An unequal crossover nogenetics 40:425. between the RCCX modules of the human MHC leading to the presence of a 22. Schneider, P. M., K. Witzel-Schlomp, C. Rittner, and L. Zhang. 2001. The en- CYP21B gene and a tenascin TNXB/TNXA-RP2 recombinant between C4A and dogenous retroviral insertion in the human complement C4 gene modulates the C4B genes in a patient with juvenile rheumatoid arthritis. Exp. Clin. Immuno- expression of homologous genes by antisense inhibition. Immunogenetics 53:1. genet. 16:81. 23. Chung, E. K., Y. Yang, C. A. Blanchong, J. F. Sotos, R. M. Rennebohm, and 47. Barba, G., C. Rittner, and P. M. Schneider. 1993. Genetic basis of human com- C. Y. Yu. 2003. Polymorphisms and rearrangements of RP-C4-CYP21-TNX plement C4A deficiency: detection of a point mutation leading to nonexpression. (RCCX) genes in major histocompatibility complex associated diseases. Pediatr. J. Clin. Invest. 91:1681. Res. 53:257A. 48. Rupert, K. L., J. M. Moulds, Y. Yang, F. C. Arnett, R. W. Warren, J. D. Reveille, 24. Reid, K. B. M., and R. R. Porter. 1981. The proteolytic activation systems of B. L. Myones, C. A. Blanchong, and C. Y. Yu. 2002. The molecular basis of complement. Annu. Rev. Biochem. 50:433. complete C4A and C4B deficiencies in a systemic lupus erythematosus (SLE) 25. Porter, R. R. 1983. Complement polymorphism, the major histocompatibility patient with homozygous C4A and C4B mutant genes. J. Immunol. 169:1570. complex and associated diseases: a speculation. Mol. Biol. Med. 1:161. 49. Frank, M. M., J. A. Gelfand, and J. P. Atkinson. 1976. Hereditary angioedema: 26. Walport, M. J. 2001. Complement: part II. N. Engl. J. Med. 344:1140. the clinical syndrome and its management. Ann. Intern. Med. 84:580. 27. Carroll, M. C. 1998. The role of complement and complement receptors in in- 50. Gompels, M. M., R. J. Lock, J. E. Morgan, J. Osborne, A. Brown, and P. Virgo. duction and regulation of immunity. Annu. Rev. Immunol. 16:545. 2002. A multicentre evaluation of the diagnostic efficiency of serological inves- 28. Yang, Y., E. K. Chung, B. Zhou, K. Lhotta, L. A. Hebert, D. J. Birmingham, tigations for C1 inhibitor deficiency. J. Clin. Pathol. 55:145. B. H. Rovin, and C. Y. Yu. 2003. The intricate role of complement C4 in human 51. Trouw, L. A., A. Roos, and M. R. Daha. 2002. Autoantibodies to complement SLE. Curr. Dir. Autoimmun. In press. components. Mol. Immunol. 38:199. 29. Atkinson, J. P., and J. A. Schifferli. 1999. Complement system and systemic 52. He, S., and Y. L. Lin. 1998. In vitro stimulation of C1s proteolytic activities by lupus erythematosus. In Lupus, Molecular and Cellular Pathogenesis. C1s-presenting autoantibodies from patients with systemic lupus erythematosus. G. M. Kammer and G. C. Tsokos, eds. Humana Press, Totowa, p. 529. J. Immunol. 160:4641. 30. Bishof, N. A., T. R. Welch, and L. S. Beischel. 1990. C4B deficiency: a risk 53. Wild, G., J. Watkins, A. M. Ward, P. Hughes, A. Hume, and N. R. Rowell. 1990. factor for bacteremia with encapsulated organisms. J. Infect. Dis. 162:248. C4a anaphylatoxin levels as an indicator of disease activity in systemic lupus erythematosus. Clin. Exp. Immunol. 80:167.

31. Jaatinen, T., O. Ruuskanen, L. Truedsson, and M.-L. Lokki. 1999. Homozygous Downloaded from deletion of the CYP21A-TNXA-RP2–C4B gene region conferring C4B defi- 54. Moulds, J. M., N. B. Warner, and F. C. Arnett. 1993. Complement component ciency associated with recurrent respiratory infections. Hum. Immunol. 60:707. C4A and C4B levels in systemic lupus erythematosus: quantitation in relation to C4 null status and disease activity. J. Rheumatol. 20:443. 32. Ellman, L., I. Green, F. Judge, and M. M. Frank. 1971. In vivo studies in C4- 55. Ho, A., S. G. Barr, L. S. Magder, and M. Petri. 2001. A decrease in complement deficient guinea pigs. J. Exp. Med. 134:162. is associated with increased renal and hemologic activity in patients with sys- 33. Bottger, E. C., T. Hoffmann, U. Hadding, and D. Bitter-Suermann. 1986. Guinea temic lupus erythematosus. Arthritis Rheum. 44:2350. pigs with inherited deficiencies of complement components C2 or C4 have char- 56. Vatay, A., Y. Yang, E. K. Chung, B. Zhou, C. A. Blanchong, M. K. Kovacs, acteristics of immune complex disease. J. Clin. Invest. 78:689. I. Karadi, G. Fust, L. Romics, L. Varga, et al. 2003. Relationship between com- 34. Finco, O., S. Li, M. Cuccia, F. S. Rosen, and M. C. Carroll. 1992. Structural

plements C4A and C4B diversities and two TNFA promoter polymorphisms in http://www.jimmunol.org/ differences between the two human complement C4 isotypes affect the humoral two healthy Caucasian populations. Hum. Immunol. 64:543. immune response. J. Exp. Med. 175:537. 57. Truedsson, L., Z. Awdeh, E. Yunis, S. Mrose, B. Moore, and C. Alper. 1989. 35. Milich, D. R. 2002. Influence of C4A deficiency on nonresponse to HBsAg vaci- Quantitative variation of C4 variant proteins associated with many MHC haplo- nation: a new immune response gene. J. Hepatol. 37:396. types. Immunogenetics 30:414. 36. Hohler, T., B. Stradmann-Bellinghausen, R. Sanger, A. Victor, C. Rittner, and 58. Dodds, A. W., X.-D. Ren, A. C. Willis, and S. K. A. Law. 1996. The reaction P. M. Schneider. 2002. C4A deficiency and nonresponse to hepatitis B vaccina- mechanism of the internal thioester in the human complement component C4. tion. J. Hepatol. 37:387. Nature 379:177. 37. De Silvestri, A., A. Pasi, M. Martinetti, C. Belloni, C. Tinelli, G. Rondini, 59. Isenman, D. E., and J. R. Young. 1984. The molecular basis for the differences L. Salvaneschi, and M. Cuccina. 2001. Family study of non-responsiveness to in immune hemolysis activity of the Chido and Rodgers isotypes of human com- hepatitis B vaccine confirms the importance of HLA class III C4A locus. Genes plement component C4. J. Immunol. 132:3019. Immun. 2:367. 60. Schifferli, J. A., G. Hauptmann, and J. Pierre-Paccaud. 1987. Complement-me- 38. Kohler, P. F., and H. J. Mu¨ller-Eberhardt. 1967. Immunochemical quantitation of diated adherence of immune complexes to human erythrocytes: difference in the by guest on September 23, 2021 the third, fourth and fifth components of human complement: concentrations of requirements for C4A and C4B. FEBS Lett. 213:415. sera in healthy adults. J. Immunol. 99:1211. 61. Gibb, A. L., A. M. Freeman, R. A. G. Smith, S. Edmonds, and E. Sim. 1993. The 39. Nelson, R. A. J., J. M. Jensen, E. Gigli, and N. Tamura. 1966. Methods for the interaction of soluble human complement recepter type 1 (sCR1, BR55730) with separation, purification and measurement of nine components of hemolytic com- human complement component C4. Biochim. Biophys. Acta 1180:313. plement in guinea pig serum. Immunochemistry 3:111. 62. Reilly, B. D., and C. Mold. 1997. Quantitative analysis of C4Ab and C4Bb 40. Miller, S. A., D. D. Dykes, and H. F. Polesky. 1988. A simple salting out pro- binding to the /C4b receptor (CR1, CD35). Clin. Exp. Immunol. 110:310. cedure for extracting DNA from human nucleated cells. Nucleic Acids Res. 16: 63. Gatenby, P. A., J. E. Barbosa, and P. J. Lachmann. 1990. Differences between 1215. C4A and C4B in the handling of immune complexes: the enhancement of CR1 41. Yu, C. Y., C. A. Blanchong, E. K. Chung, K. L. Rupert, Y. Yang, Z. Yang, binding is more important than the inhibition of immunoprecipitation. Clin. Exp. B. Zhou, and J. M. Moulds. 2002. Molecular genetic analyses of human com- Immunol. 79:158. plement components C4A and C4B. In Manuals of Clinical Laboratory Immu- 64. Reilly, B. D. 1999. Analysis of human C4A and C4B binding to an immune nology, 6th Ed. N. R. Rose, R. G. Hamilton, and B. Detrick, eds. ASM Press, complex in serum. Clin. Exp. Immunol. 117:12. Washington DC, p. 117. 65. Paul, L., V. M. Skanes, J. Mayden, and R. P. Levine. 1988. C4-mediated inhi- 42. Awdeh, Z. L., D. Raum, and C. A. Alper. 1979. Genetic polymorphism of human bition of immune precipitation and differences in inhibitory action of genetic complement C4 and detection of heterozygotes. Nature 282:205. variants, C4A3 and C4B1. Complement 5:110. 43. Sim, E., and S. Cross. 1986. Phenotyping of human complement component C4, 66. Varga, L., C. A. Alper, Z. Zam, and G. Fust. 1991. Decreased inhibition of a class III HLA antigen. Biochem. J. 239:763. immune precipitation by sera with the C2 B allotype. Clin. Immunol. Immuno- 44. Yu, C. Y., K. T. Belt, C. M. Giles, R. D. Campbell, and R. R. Porter. 1986. pathol. 59:65. Structural basis of the polymorphism of human complement component C4A and 67. Holme, E., J. Veitch, A. Johnston, I. McKay, and K. Whaley. 1992. The relative C4B: gene size, reactivity and antigenicity. EMBO J. 5:2873. roles of C4A and C4B in prevention of immune precipitation, solubilization and 45. Chung, E. K., Y. Yang, K. L. Rupert, K. N. Jones, R. M. Rennebohm, . Rheumatol. Int. 12:187. C. A. Blanchong, and C. Y. Yu. 2002. Determining the one, two, three or four 68. Klint, C., L. Truedsson, and G. Sturfelt. 1995. Binding to erythrocyte comple- long and short loci of human complement C4 in a major histocompatibility com- ment receptor type 1 of BSA/anti-BSA complexes opsonized by C4A3 or C4B1 plex haplotype encoding for C4A or C4B proteins. Am. J. Hum. Genet. 71:810. in the presence of serum. Scand. J. Immunol. 42:425.