CMAH genotyping survey for blood types A, B and C (AB) in purpose-bred cats

Summary

In domestic cats, the AB blood group system consists of the three types A, B and C (also called AB). Mismatches can cause acute hemolytic transfusion reactions and hemolysis of the newborn (neonatal isoerythrolysis, NI). As blood types B and C are inherited recessively to A, breeders need to know the genotype to predict blood types in offspring and avoid NI. Several CMAH variants have been described as being associated with the b and a^c alleles, and different genotyping schemes exist. Here, we genotyped 2145 cats with the original SNV panel, including SNVs c.142G>A and Δ−53, and our new scheme, with SNVs c.179G>T, c.268T>A and c.1322delT, to differentiate types A and B and added the SNV for the common a^c (c.364C>T). Based upon the new scheme, all samples were assigned the correct genotype. No discordances appeared for the A allele, and new breed-specific SNVs (c.179G>T, c.1322delT) for the b allele were discovered. Furthermore, the genotypes A/a^c (type A), a^c/a^c (C) and a^c/b (C) could be detected. We found the variant c.179G>T in additional breeds: Ragdoll, Siberian, Scottish Fold, Chartreux, Neva Masquerade, British Shorthair and Highlander. Also, the variant c.364C>T was detected in additional breeds: Bengal, British Shorthair, Maine Coon, and Scottish Fold. We conclude that our new SNV panel is superior in genotyping cats than the original SNV panel and assures correct assignments of types A, B and C to assist veterinary clinicians and breeders to recognize, confirm and avoid blood incompatibilities such as acute hemolytic transfusion reactions and NI.

The feline AB blood group system is composed of type A, B and C (also called AB). It is of great clinical importance as blood incompatibilities can result in acute hemolytic transfusion reactions and hemolysis of the newborn (neonatal isoerythrolysis [NI]) due to presence of naturally occurring anti-A alloantibodies (Auer & Bell 1981; Auer et al. 1982; Giger et al. 1991; Giger 2014). The enzyme cytidine monophosphate-N-acetylneuraminic acid hydroxylase (CMAH; EC 1.14.18.2) converts sialic acid N-acetylneuraminic acid (NeuAc; type B antigen) to N-glycolylneuraminic acid (NeuGc; type A antigen) (Bighignoli et al. 2007).

Although immunohematological (serological) blood typing is clinically useful, genotyping cats is important in breeding purpose-bred cats to predict blood types in offspring and avoid NI. The CMAH locus is proposed to have three alleles with several haplotypes in the allelic series of A>a^c>b (Giger et al. 1991; Griot-Wenk & Giger 1995). Thus, genotypes A/A, A/b and A/a^c can be observed in type A cats, b/b in type B and a^c/a^c and a^c/b in type C cats.

Many DNA polymorphisms have been described in the cat CMAH gene: Δ−53, c.139C>T, c.142G>A, c.179G>T, c.187A>G, c.268T>A, c.327A>C, c.364C>T, c.374C>T, c.376G>A, c.593A>C, c.868A>C, c.898A>G, c.933delA, c.1322delT, c.1342G>A and c.1603G>A. Several of these variants are presumed to lead to loss or reduction of the regular CMAH activity needed for the type A antigen and thereby lead to blood type B and C respectively (Bighignoli et al. 2007; Tasker et al, 2014; Gandolfi et al. 2016; Omi et al. 2016; Kehl et al. 2018). Originally SNVs c.142G>A and Δ−53 were used in diagnostic genotype differentiation between type A (genotype A/A and A/b) and type B (genotype b/b) (Bighignoli et al. 2007), but it was recognized that there were many genotype–phenotype discordances (Kehl et al. 2018). We recently showed that the CMAH SNV c.268T>A (p.Tyr90Asn) instead of the above was most likely to cause enzyme dysfunction. Indeed, this SNV exhibited perfect genotype–phenotype correlation in a limited number (n = 421) of cats (Kehl et al. 2018). Additionally, the SNVs c.179G>T (p.Gly60Val) and c.1322delT (p.Leu441*) are detrimental to CMAH function and were shown to be associated with the b allele in our initial genotyping–phenotyping survey (Kehl et al. 2018). Also, we and others have recently associated the SNV c.364C>T (p.Pro122Ser) with type C in Ragdolls (Gandolfi et al. 2016, Kehl et al. 2018). We therefore proposed a new genotyping scheme including these four SNVs (Kehl et al. 2018) (Table 1). Here we compare the genotyping results from the original to the new scheme in a large group of cats to further elucidate the relevance of the different SNVs in routine blood typing.

Table 1.

Comparison of genotyping and predicted phenotypic blood typing results according to the original and new SNV panel of 2145 purpose-bred cats.

New SNV panel					Original SNV panel
Genotype	c.179G>Ta	c.268T>A2	c.364C>T3	c.1322delT4	c.142G>A2	Δ−53b	Genotype		Number of cats	Predicted phenotype
A/A	GG	TT	CC	TT	GG	NN	A/A		1290	A (91.7%)
	GG	TT	CC	TT	GG	NP	Discordant		93
	GG	TT	CC	TT	GG	PP	Discordant		2
A/b	GG GT GG	TA TT TT	CC CC CC	TT TT T*	GA GG GG	NP NN NN	A/b A/A A/A		488 15 37
	GG	TA	CC	TT	GG	NP	Discordant		3
	GG	TA	CC	TT	AA	NP	Discordant		1
A/ac	GG	TT	CT	TT	GG	NN	A/A		37
	GG	TT	CT	TT	GG	NP	Discordant		1
	GG	TT	CT	TT	GA	NN	Discordant		1
ac/b	GG	TA	CT	TT	GA	NP	A/b		4	C (0.7%)
	GG	TT	CT	T*	GG	NN	A/A		8
ac / ac	GG	TT	TT	TT	GG	NN	A/A		3
b/b	GG	AA	CC	TT	AA	PP	b/b		145	B (7.6%)
	TT	TT	CC	TT					x
	GG	TT	CC	**	GG	NN	A/A		7
	GT	TA	CC	TT	AA	NP	Discordant		4
	GG	TA	CC	T*	GA	NP	A/b		6
	GT	TT	CC	T*					x

N not present, P present;

^*deletion, X not seen.

¹Omi et al. 2016;

²Bighignoli et al. 2007; Gandolfi et al. 2016

^dKehl et al. 2018.

All ethylenediaminetetraacetic blood and cheek swab samples from purpose-bred cats were submitted to Laboklin by veterinary clinicians and breeders for routine diagnostic genetic testing from November 2017 to October 2018. Most samples originated from Germany, but samples were also received from other countries throughout Europe. Genomic DNA was isolated (MagNAPure 96; Hoffmann-La Roche), and genotyping for the CMAH SNVs Δ−53, c.142G>A, c.179G>T, c.268T>A, c.364C>T and c.1322delT was performed (TaqMan SNP Assays; ThermoFisher), using FastStart Essential DNA Probes Master and LightCycler 480 II (Hoffmann-La Roche), as previously described (Kehl et al. 2018). Routine immunological blood typing was performed for 37 cats using an immunochromatographic strip method, according to the manufacturer’s instructions [Alvedia rapid-test (LabTest A+B)]. Results and correlation between SNVs, genotypes and predicted phenotypes (with a few also regularly typed cats) were evaluated.

Our survey included 2145 pedigreed cats from 31 breeds, among which there were five breeds with more than 100 cats genotyped and a few breeds with fewer than 10 cats genotyped (Table 2). Compared to prior phenotypic blood typing surveys of different breeds (Giger et al. 1991; Jensen et al. 1994; Knottenbelt et al. 1999; Weingart et al. 2006; Forcada et al. 2007; Spada et al. 2014; Vieira et al. 2017), equal or more purpose-bred cats had type A blood, which may be due to breeders’ selection for type A cats and the breeders’ interest in determining if their type A cats are carriers for the b or a^c allele.

Table 2:

Distribution of CMAH genotypes of 2145 purpose-bred cats from 31 breeds (16 breeds with n ≥ 8 cats per breed shown; additional 14 breeds shown in footnote)

	Genotypes
Breed	A/A	A/b	b/b	b/b compound1	A/ac	ac/b	ac/ac	n
Abyssinian	25	6	1					32
Bengal	136	2			9		1	148
Birman	55	49	2					106
British Shorthair	128	189	109	1	2			429
Chartreux	2	3	2	1				8
Devon Rex	6	9	3					18
Highlander	13	17	7	1				38
Maine Coon	787	159	10		2			958
Neva Masquerade	4	7		1				12
Norwegian Forest	45	3						48
Persian	11	3	1					15
Ragdoll	90	61	9	6	24	12	2	204
Savannah	9	1						10
Scottish Fold	6	6	4		1			17
Siberian	17	8	2					27
Somali	6	8	1					15

Note: Australian Mist (n = 1), Burma (n = 2), Cornish Rex (n = 2), Exotic Shorthair (n = 6), Oriental Shorthair (n = 6), Russian Blue (n = 6), Siamese (n = 2) and Tonkinese (n = 2) all had the regular A/A genotype (all type A blood). Egyptian Mau (n = 6), Ocicat (n = 3), Savannah (n = 10), Sphynx (n = 4), Thai (n = 5) and Turkish Angora (n = 5) were either A/A or A/b (all type A). In addition, one Thai cat was (A/a^c) which still is a type A cat.

¹b/b compound cats are cats with heterozygous genotypes at two loci in CMAH gene (c.179G>T, c.268T>A and c.1322delT) suggesting the genotype b/b.

Overall 162 (7.6%) cats had the genotype b/b (type B blood) and were homozygous for the A allele at position 268 (n = 145) or for the deletion at position 1322 (n = 7) or compound heterozygous for the SNVs c.268T>A/c.179G>T (n = 4) or c.268T>A/c.1322delT (n = 6). Thus, genotyping for this additional b allele is important.

The variant c.364C>T was described as being the cause of type C in Ragdolls and a few random-bred cats (Gandolfi et al. 2016; Kehl et al. 2018). Of the 204 Ragdolls tested, indeed 24 (12%) showed the genotype A/a^c (type A), 12 (6%) had the genotype a^c/b (type C) and two (1%) were homozygous a^c/a^c (type C). Among the 1941 genotyped non-Ragdoll cats, the variant c.364C>T was rarely found: in only nine heterozygous and one homozygous Bengals, two British Shorthairs, two Maine Coons and one Scottish Fold cat (and one Thai-Siam-hybrid) in the heterozygous state (expressing type A blood). The homozygous a^c/a^c Bengal cat was confirmed to possess type C blood by phenotyping. Thus, the variant c.364C>T seems to be more widespread and exist in more breeds than originally described (Gandolfi et al. 2016; Kehl et al. 2018). Additionally, more than one genetic cause for type C can be suspected in other breeds, as one British Shorthair cat with type C (and without the c.364 variant) has been described (Gandolfi et al. 2016).

The variant c.179G>T was recently described as causative for type B in Turkish Angora, Neva Masquerade as well as Domestic Shorthair cats (Kehl et al. 2018). In this study, this variant was also found in five Ragdolls, four Siberian cats and one Scottish Fold cat in the heterozygous state together with an A allele and thus type A blood. Additionally, the variant was found in one British Shorthair, Highlander, Chartreux and Neva Masquerade cats in the heterozygous state together with the variant c.268T>A, suggesting a compound heterozygous status causing type B blood. However, their blood types were not confirmed by immunohematological testing.

The variant c.1322delT was recently described as being causative for type B in Ragdoll and Scottish Fold cats (Kehl et al. 2018). In the present study, this variant was found exclusively in Ragdolls: 54 cats (23%) were heterozygous for the c.1322delT variant; six of these 54 cats were also heterozygous for c.268T>A, suggesting the genotype b/b (type B); and eight of these 54 cats showed the genotype a^c/b (type C). If these two variants c.179G>T and c.1322delT would not have been included in the genotyping scheme, 77 (3.6%) cats would have been genotyped incorrectly.

Furthermore, due to contradicting diplotypes for c.142G>A and Δ−53, 105 (5%) of genotyped cats could not be assigned to one blood type with the original SNV panel. Noteworthy, these discordances were found in different breeds, but especially in Bengals (10.3% with genotype A/A) and Maine Coons (8.8% with genotype A/A). However, based upon our new SNV panel, these 105 cats could be readily assigned as having type A or B blood (Table 1).

Additionally, we assigned different genotypes to 54 cats compared to the original SNV panel: 39 type A, A/a^c; three type C, a^c/a^c; and 12 type C, a^c/b. Of the 39 A/a^c cats, 37 would have been genotyped as A/A with the original scheme and the other two cats already had discordant genotyping results with the original two SNVs. All three a^c/a^c cats and eight of the 12 a^c/b cats would have been genotyped as A/A and four as A/b with the original SNV panel. The original scheme also missed 10 type B cats that were b/b compound heterozygotes and seven type B cats homozygous for the SNV c.1322delT.

All A/A genotyping results of the new scheme were typed A/A in the original genotyping scheme, whereas 95 cats with discordant diplotypes for the SNVs c.142G>A and Δ−53 were assigned A/A. Based upon our new SNV panel all cats with undetermined blood type could be assigned a specific blood type.

The 37 cats that we also blood typed by an immunohematological method showed perfect genotype–phenotype correlation, further supporting the excellent genotype–phenotype correlation shown in our prior study of 421 cats (Kehl et al. 2018).

This survey shows the superiority of the new genotyping scheme (SNVs c.179G>T, c.268T>A, c.364C>T and c.1322delT) over the old scheme (SNVs c.142G>A and Δ−53). Type C cats with the genotypes a^c/a^c and a^c/b can now also be detected. Additionally, the B type caused by either SNVs c.179G>T or c.1322delT alone or as a compound heterozygote with the SNV c.268T>A can be found. Moreover, no genotyping–phenotyping discordances were observed. The new SNV panel also demonstrates its strength in detecting the ‘breed-related’ SNVs, such as c.179G>T, c.364C>T and c.1322delT, in additional breeds.

In conclusion, the new genotyping scheme for the feline AB blood group system applied in this large survey provides detailed typing results that are important for breeders and veterinarians confronted with inconsistent immunohematological typing results. Although genotyping is not yet a replacement for immunohematological testing in feline clinical practice, this new genotyping scheme is an additional step forward to assuring blood compatibility in cats, as already used in human blood banking. There may well be additional new SNVs discovered in some breeds and geographical regions in the future that may cause type B and C and may have to be added to the current genotyping scheme to be inclusive.