1st Page

Journal of Animal Science - Animal Genetics

Frequency of leukochimerism in Holstein and Jersey twinsets12


This article in JAS

  1. Vol. 94 No. 11, p. 4507-4515
    Received: May 12, 2016
    Accepted: Aug 10, 2016
    Published: October 7, 2016

    3 Corresponding author(s):

  1. A. S. Younga and
  2. B. W. Kirkpatrick 3a
  1. a University of Wisconsin–Madison, Madison 53706


The current study was conducted to test breed difference in the frequency of leukochimerism. This study used leukochimerism as evidence of placental vascular anastomosis formation and compared its frequency in the Holstein and Jersey breeds. We test the null hypothesis that there is no difference in incidence of leukochimerism in the Holstein and Jersey breeds. Hair and blood samples were collected from 85 Jersey twinsets and 80 Holstein twinsets, ranging in age from 1 d to 8 yr. An additional 7 Holstein twinsets (6 complete and 1 partial where 1 twin died) were sampled originally 48 to 72 h after birth and resampled at 5 to 10 mo of age to provide an assessment of whether leukochimerism changed with age. DNA was extracted from white blood cells (potentially chimeric) and hair follicles (not chimeric). DNA samples were successfully genotyped for 19 SNP selected for high minor allele frequency in both breeds based on previous bovine 50K genotyping. The genotyping assays provided quantitative data that was used to assess chimerism in blood-derived DNA. Monozygotic twins, as a percentage of all twin births, were 3.5 and 9.1% for the Jersey and Holstein breeds, respectively. Jersey and Holstein breeds did not differ in proportion of nonchimeric twinsets at 20.1 and 15.7%, respectively (P > 0.05), providing no evidence for genetic variation in anastomosis. The degree of chimerism for members of a twinset was also evaluated with regard to representation of self vs. co-twin in the blood-derived DNA. For twinsets where the more chimeric twin was 45% or greater co-twin in its blood-derived DNA, there was a strong inverse relationship (P < 0.001) between percent co-twin in the blood-derived DNA of members of a twinset. For twinsets where the more chimeric twin was less than 45% co-twin, there was no significant relationship between the degrees of chimerism in members of the twinset. These results suggest that variation in chimerism in members of a twinset may be a function of degree of anastomosis and differences in timing of the migration of hematopoietic stem cells between members of the twinset.


Placental vascular anastomosis (PVA) occurs between the placentas of the developing twins when blood vessels of both placentas fuse, allowing blood and all its components, including white blood cells and their predecessor stem cells, to flow between fetuses resulting in animals that are leukochimeric. One of the most researched consequences of PVA is freemartinism. Freemartins are females born co-twin to a male, who have abnormal development of their reproductive tract due to the presence of the anti-Mullerian hormone (AMH) produced by the male co-twin. The Sertoli cells of the male’s developing testis secrete AMH, which affects development of the Mullerian duct, preventing normal formation of the uterus and oviducts. Approximately 82.5 to 95.7% of females born co-twin to a male will be freemartins (Zhang et al., 1994; Gregory et al., 1996). While freemartinism is the most notable outcome of PVA, another aspect is that the formation of anastomoses between fetuses means that they are no longer independent. This is evidenced by the low survival rate of a single embryo following embryo reduction of a twinset of approximately 5% (López-Gatius and Hunter, 2005).

This study tests the null hypothesis that the proportion of leukochimeric twinsets in the Jersey breed are greater than or equal to the Holstein breed against a 1-sided alternative hypothesis that the incidence of leukochimerism in the Jersey is lower vs. the Holstein. This alternative hypothesis was suggested by results from a study looking at interspecific chimerism in embryo transfer-derived twinsets, where Jersey-Brahman twinsets had significantly lower incidence of blood cell chimerism than Friesian- Brahman twinsets (Summers et al., 1984). Embryonic or fetal loss of members of a twinset is not independent (Bennett et al., 1998), potentially due to the formation of anastomoses and a shared placental blood supply. In a twinning context, selection against formation of anastomoses may be desirable.


The University of Wisconsin–Madison College of Agricultural and Life Sciences Animal Care and Use Committee approved this research.


Blood and hair samples for the study were obtained from animals in commercial, university, and federal dairy operations. Samples for the Holstein population were obtained from 1 commercial Wisconsin dairy, the University of Wisconsin–Madison’s dairy research herd, and the USDA Dairy Forage Research Station herd. Samples for the Jersey population were obtained from 5 commercial dairies that were located in Wisconsin (n = 2), Iowa (n = 2), and Minnesota. Animals were identified as members of intact twinsets (both twins living) using herd records. Birth type was not routinely recorded, so members of a twinset were identified by having a common sire, dam, and date of birth; animals with nonsequential identification numbers were removed from consideration. Twinsets were biased toward female-female sets due to the typical practice of removing males and potentially freemartin females from the herd at a young age. Hair samples were obtained by pulling hairs from the tail or tail switch if present and contained the hair follicle. Blood samples were collected in 10-mL Covidien Monoject (Covidien, Dublin, Ireland) blood collection tubes containing 0.10 mL of a 15% EDTA (K3) solution from the coccygeal vein or jugular vein. Samples from 85 Jersey twinsets and 80 Holstein twinsets were collected. An additional 7 Holstein twinsets (6 complete and 1 partial where 1 twin died), originally sampled 48 to 72 h after birth, were resampled at ages ranging from 3 to 12 mo to provide an assessment of whether leukochimerism changed with age.

DNA Preparation

Upon collection, samples were immediately placed on ice for transportation to the laboratory. Hair samples were stored at −20°C, and blood samples were spun at 1,500 × g at 4°C for 10 min, after which the buffy coat was transferred to a 1.5-mL Eppendorf tube and stored at −20°C. The DNA extraction method used was a modification of a previously published protocol (Cruickshank et al., 2004) with 2-mercaptoethanol used in place of dithiothreitol. DNA purity was assessed with a spectrophotometer, and samples that had an A260/A280 nanometer ratio below 1.70 were reextracted starting from the DNA sample using the ReliaPrep Blood gDNA Miniprep System (Promega Corp., Madison, WI).


Twenty SNP were selected for use based on high minor allele frequencies (ranging from 0.39 to 0.50 in both Holstein and Jersey breeds), high call rates (>0.97) in previous Bovine SNP50 chip assays based on our previous experience (Table 1; Alpay et al., 2014; Zare et al., 2014), and genomic location (all SNP were from different chromosomes). SNP genotyping was performed by LGC Genomics (Beverly, MA) using KASP assays. KASP genotyping assays are based on competitive allele-specific PCR. Two unique allele-specific forward primers and a common reverse primer are used to amplify the targeted DNA fragment. The allele-specific primers each contain a unique tail sequence that corresponds with a universal fluorescence resonant energy transfer (FRET) cassette. After PCR, endpoint fluorescence was read to quantify alternative allele products. Alternative alleles were labeled with fluorescein (Fam) and hexachlorofluorescein (Hex) fluorescence dyes, and a carboxy-X-rhodamine (Rox) passive reference dye was coloaded with each sample to provide a control for variation in sample loading. Fam and Hex fluorescence intensities were adjusted by dividing by the control Rox fluorescence but were not otherwise normalized. SNP BTB-00921095 was not used in subsequent analyses, as no heterozygotes were observed, indicating assay failure. Data were used to calculate a relative intensity value for 1 of the 2 SNP alleles (denoted as allele 1). Using only the hair-derived DNA samples, a range of observed values for nonchimeric samples was determined for each genotype category of each SNP by breed. Observed relative intensity values for blood-derived DNA samples were compared against this range to assess whether or not a sample was considered chimeric as described in more detail below. Recognizing that homozygote genotypes could reasonably have relative intensity scores of up to 1 for allele 1 in the case of homozygotes for allele 1 (i.e., 100%) and a minimum value of 0 for allele 1 in the case of homozygotes for allele 2, the maximum threshold for allele 1 was expanded to 1 and the minimum threshold for allele 2 was expanded to 0. A t test was performed to test a difference in mean relative intensity between breeds for each genotype within each SNP. Seventeen out of 19 SNP were not significant at a P value of 0.01 for all genotypes, and the data for the 2 breeds were combined and the minimum and maximum values for the 17 SNP were recalculated to include both breeds.

View Full Table | Close Full ViewTable 1.

SNP selected for the use of genotyping with both the Holstein and Jersey DNA samples

SNP name BTA1 Call rate2 MAF3 Call rate2 MAF3
ARS-BFGL-NGS-12933 1 0.992 0.412 0.979 0.457
BTB-01601989 2 0.989 0.430 0.980 0.497
BTB-01327945 4 0.989 0.437 0.982 0.401
BTB-01133499 6 0.987 0.442 0.979 0.471
Hapmap43946-BTA-80204 7 0.989 0.488 0.987 0.412
Hapmap33532-BTA-84282 9 0.992 0.484 0.980 0.483
Hapmap23524-BTC-065402 14 0.987 0.444 0.981 0.477
ARS-BFGL-BAC-35394 15 0.984 0.469 0.984 0.451
ARS-BFGL-NGS-1152 16 0.989 0.445 0.981 0.489
BTB-00682000 17 0.981 0.424 0.980 0.425
Hapmap51593-BTA-42524 18 0.995 0.450 0.980 0.478
Hapmap44079-BTA-45543 19 0.981 0.440 0.979 0.476
ARS-BFGL-NGS-110086 20 0.992 0.488 0.980 0.409
BTB-01303828 21 0.981 0.452 0.980 0.449
Hapmap42246-BTA-54605 22 0.989 0.481 0.979 0.457
ARS-BFGL-NGS-89490 23 0.995 0.433 0.980 0.473
Hapmap41303-BTA-58461 24 1.000 0.495 0.981 0.416
Hapmap30460-BTC-059446 25 0.992 0.438 0.979 0.449
BTB-00921095 26 0.992 0.391 0.980 0.478
ARS-BFGL-NGS-56644 27 0.981 0.456 0.978 0.497
1BTA = Bos taurus chromosome.
2Proportion of samples successfully genotyped for the SNP in previous studies (Alpay et al. 2014; Zare et al. 2014).
3MAF = minor allele frequency observed in previous studies (Alpay et al. 2014; Zare et al. 2014).

Determining Leukochimerism

Each individual’s hair and blood-derived DNA of a twinset was then compared with its counterpart. Monozygotic twinsets were identified by the twins having the same genotype for all 19 SNP based on genotypes derived from hair-derived DNA. When the genotypes from hair-derived DNA of each twin differed at a specific SNP (an informative SNP), the relative intensity value of allele 1 from genotyping blood-derived DNA was compared against previously observed ranges for the 3 genotypes of that SNP. A relative intensity value outside the cluster previously observed for the expected genotype (based on the individual’s hair-derived DNA sample) was interpreted as evidence of chimerism for that SNP for that individual (Fig. 1). Identification of putative chimerism was also used for quality control purposes; twins with identical genotypes based on hair-derived DNA but chimerism as evidenced by blood-derived DNA were considered potentially contaminated samples and discarded from subsequent analyses. Three Holstein twinsets were excluded for this reason.

Figure 1.
Figure 1.

Illustration of interpretation of evidence for chimerism. The histogram shows the frequency of relative intensity values for allele 1 using only hair-derived DNA from the combined Holstein and Jersey populations for 1 SNP, Hapmap51593-BTA-42524. The relative intensity values cluster for the 3 genotypes, TT, CT, and CC. Solid blue lines depict the range of relative intensity values observed for each genotype. Allele 1 (allele 2) homozygotes could logically have relative intensity scores as great as 1 (as low as 0) for allele 1, thus the maximum (minimum) acceptable value for relative frequency of allele 1 (allele 2) was extended to 1 (0) (gray line). Representative chimeric (32382) and nonchimeric (32383) twins are indicated by red squares. Blood-derived DNA relative intensity value for 32382 is inconsistent with its hair-derived DNA genotype of TC, indicating chimerism. In contrast, blood-derived relative intensity value for 32383 is consistent with its hair-derived DNA genotype of TT, and its classification was nonchimeric.


The proportion of co-twin influence in each animal’s blood-derived DNA was calculated at all informative SNP. A unique standard curve was determined for each SNP based on simple linear regression. Genotype was the dependent variable and was coded as 1 for homozygous for allele 1, 0.5 for heterozygotes, and 0 for homozygotes for allele 2. Genotype was regressed on relative intensity value using hair-derived DNA data. The relative intensity value from blood-derived DNA was then used in the standard curve to estimate a continuous genotypic value characterizing the genotypic mix in chimeric twins. Proportion of co-twin influence in each animal’s blood-derived DNA was calculated as follows:where Bs is the continuous genotypic value from blood-derived DNA from self, estimated from SNP-specific standard curves, and HS and HC are discontinuous genotypic values from hair-derived DNA from self (HS) or co-twin (HC), coded as 1, 0.5, or 0 for homozygotes for allele 1, heterozygotes, or homozygotes for allele 2, respectively.

Proportion of co-twin influence was then averaged across all informative SNP. In cases where the hair-derived genotype at a SNP for an animal or its co-twin was unknown, no calculation was made.

Statistical Analysis

Differences between breeds in incidences of leukochimeric twinsets were analyzed using Proc GLIMMX in SAS (SAS Inst. Inc., Cary, NC) with a linear model, including the fixed effect of breed and the random effect of herd within breed, with a logit link since the response variable is binomial. Breed effect was tested using herd within breed as the error term. The relationship between the proportions of co-twin influence of twins was analyzed by linear regression using Proc GLM in SAS (SAS Inst. Inc.), with the twin with the greater proportion of co-twin influence as the dependent variable and the twin with the lesser proportion as the independent variable. Effects of breed and herd within breed were not significant (P > 0.05) and were not included in the final model.


Genotype distribution of relative intensity values was assessed for each SNP (Fig. 2). The SNP ARS-BFGL-NGS-89490 and BTB-01133499 were significantly different between breeds (P < 0.01), and subsequent analyses of chimerism were done within breed for these SNP. Relative intensity distribution for all other SNP did not differ by breed, and chimerism was analyzed using pooled breed data. The number of informative SNP differed for each twinset, ranging from 2 to 17 with a mean of 9.05 ± 3.95 (SD; Fig. 3). The distribution of the proportion of informative SNP, for which a twin was chimeric, skewed toward a bimodal distribution with peaks at 0 and near 1 (Fig. 4).

Figure 2.
Figure 2.

Distribution of genotype data by relative intensity value. Jersey and Holstein breeds were combined for 17 SNP, as differences in mean relative intensity between breeds for each genotype were not significant (P > 0.05). Distribution differed (P < 0.05) for 2 SNP (ARS-BFGL-NGS-89490 and BTB-01133499), and they were not combined (depicted separately). The x axis indicates the hair-derived DNA relative intensity value, and the y axis is the percentage of samples with a given relative intensity value. SNP BTB-00921095 was discarded as no heterozygote genotypes were observed, suggesting assay failure.

Figure 3.
Figure 3.

Distribution of the number of SNP that are informative across twinsets. The 10 twinsets that have 0 informative SNP correspond to 3 Jersey and 7 Holstein twinsets presumed to be monozygotic. The number of informative SNP per twinset ranges from 2 to 17 out of a possible 19.

Figure 4.
Figure 4.

Distribution of percentage of informative SNP determined to be chimeric for each dizygotic twin for the Holstein and Jersey populations combined. Chimerism was determined using the relative intensity value for allele 1 derived from genotyping blood-derived DNA in comparison.


Based on the herd records, frequency of twin birth varied between 4.2 and 5.0% in the sampled Holstein herds and 1.2 and 3.6% in the 5 Jersey herds. Among the genotyped twinsets, 3 of 85 Jersey twinsets and 7 of 77 Holstein twinsets were monozygotic (Fig. 3). The proportion of observed monozygotic twinsets did not differ between breeds (P > 0.05).

Among dizygotic twinsets, for 17 out of 82 (20.7%) Jersey twinsets, both twins were nonchimeric. Similarly, for 11 out of 70 (15.1%) Holstein twinsets, both twins were nonchimeric. The proportions of nonchimeric twinsets did not differ significantly between breeds (P > 0.05). Fourteen out of the 82 Jersey twinsets (17.1%) and 10 out of the 70 Holstein twinsets (14.3%) had 1 twin for which no SNP were classified as chimeric while its co-twin had at least 2 chimeric SNP. Twinsets were classified as chimeric if either 1 or both twins exhibited chimeric SNP.

Of 7 twinsets sampled at 2 ages, 1 set was monozygotic and 6 were dizygotic (see Supplemental Table online). One dizygotic twinset showed signs of DNA contamination and was removed from further analysis. Of the remaining 5 twinsets, for 3, both twins were identified as chimeric. The living twin of the partial twinset was found to be chimeric, and the remaining twinset was chimeric, with 1 twin chimeric and 1 twin nonchimeric. All twins identified as chimeric at 2 d of age were similarly identified as chimeric at a later age, ranging from 3 to 12 mo.

The proportion of the co-twin’s influence on leukochimerism was calculated for all twins from dizygotic, chimeric twinsets (Fig. 5). Data are plotted by twinset, with the twin with a lesser percentage of co-twin influence vs. the twin with the greater percentage of co-twin influence. Distribution of the data suggests a different relationship for twinsets with more extensive leukochimerism (more chimeric twin of twinset > 45% co-twin) vs. less extensive leukochimerism (more chimeric twin of twinset ≤ 45% co-twin), with the former having a highly significant (P < 0.001) inverse relationship in proportion to the co-twin influence vs. no association for the latter (P > 0.05).

Figure 5.
Figure 5.

Relationship between the proportions of co-twin influences between members of a twinset. Each point represents a twinset, with the x axis corresponding to the twin with greater leukochimerism and the y axis corresponding to the twin with lesser leukochimerism. An arbitrary threshold (45% co-twin influence in the twin with greater leukochimerism) was used to divide the data into low (black diamonds) and high (blue circles) relative levels of leukochimerism. Twinsets with lesser leukochimerism did not exhibit a significant (P > 0.05) relationship between co-twin influences in members of the twinset (solid black line). In contrast, twinsets with greater leukochimerism had a strong inverse relationship between co-twin influences for members of the twinset (P < 0.001; solid blue line).



The placenta is the transient organ of metabolic interchange between the conceptus and the dam. Cattle have what is known as a cotyledonary placenta, which has a point of interface between the conceptus and dam known as the placentome, which consists of the fetal cotyledon and the maternal caruncle. The bovine cotyledonary placenta contains 70 to 120 cotyledons (Senger 1997), which are button-like structures of trophoblastic origin consisting of abundant blood vessels and connective tissues. At 33 d after ovulation, a fragile attachment between the maternal and chorionic tissue occurs at 3 to 4 caruncles immediately surrounding the embryo. By d 35, the attachment is stabilized and the embryo is starting to receive nutrients through the cotyledons (Melton et al., 1951). Lillie (1917) showed that fusion of the chorions usually occurs in cattle twins, followed by anastomosis of the blood vessels of the placenta. Leukocyte chimerism or leukochimerism in twins is the result of vascular anastomoses forming between the placentas of the developing twins, allowing the exchange of hematopoietic stem cells between the twins. Hence, the leukocyte population of each twin is potentially derived from a mix of leukocytes from both twins.

Placentation includes extensive angiogenesis in maternal and fetal placental tissues, accompanied by a marked increase in uterine and umbilical blood flow (Reynolds and Redmer, 1995). Angiogenesis, which refers to the formation of new vascular beds, is an important aspect of placental growth and attachment (Reynolds and Redmer, 2001). A knockout study of vascular endothelial growth factor (VEGF) receptors FLT-1 and FLK-1 in mice resulted in defects in fetal and placental vasculogenesis and angiogenesis, resulting in embryonic death around d 8 of a 20 d gestation (Fong et al., 1995). A second knockout study of VEGF in mice led to significant cardiovascular and extraembryonic vasculature defects and was lethal around d 11 of 20 (Carmeliet et al., 1996; Ferrara et al., 1996). The fibroblast growth factor (FGF) family has also been shown to be an autocrine growth factor, controlling the proliferation, migration, and differentiation of tissues of the vascular endothelial cells (Gospodarowicz, 1991). The VEGF and FGF genes both have been shown to be major angiogenic growth factors of the placenta and may play a role in the formation of placental vascular anastomoses.

During mammalian development, stem cells sequentially occupy the embryonic yolk sac, fetal liver, spleen, and adult bone marrow (O’Donoghue and Fisk, 2004). The hematopoietic stem cells that are passed from the fetal liver to the fetal spleen and finally to the bone marrow are the progenitors that will produce all other blood cells (erythrocytes, leukocytes, and platelets). These hematopoietic stem cells are self-renewing, and once in the bone marrow, they can differentiate into a variety of specialized cells and can be mobilized out of the bone marrow into the circulating blood. The migration of these stem cells allows the exchange of hematopoietic stem cells between co-twins through the placental vascular anastomoses and allows a mixed population of hematopoietic stem cells to become established in each individual of a twinset.

While the migration of cells during early development allows for the formation of mixed populations, cells like skin and hair follicles are not expected to have DNA derived from a mixed population. However, Plante et al. (1992) showed evidence of chimerism in cultured fibroblasts, indicating that some chimerism occurs in the skin albeit at a low level. Analyses in the current study were conducted assuming that the chimerism of hair follicle-derived DNA was negligible. If low levels of chimerism were present, it would result in greater variation in the relative intensity values within genotypes, leading to a more conservative calling of chimerism, or conversely, a more liberal calling of nonchimerism.

Selection of SNP markers in this study was effective, resulting in an average of 9.05 informative SNP per twinset. There was variation between SNP in the range of relative intensity values for alternative genotypes. The most suitable markers are those with the narrowest ranges of relative intensity values. For future work, it would be useful to use a larger panel of SNP permitting identification of SNP with tighter genotype distributions. Additionally, with a larger panel of SNP, analysis could be restricted to those markers for which twins had alternative homozygous genotypes. Both factors would improve the calling of chimerism and nonchimerism.

Cattle, a monotocous species, normally ovulate 1 ovum per cycle and give birth to 1 calf per year. However, on occasion, the reproductive systems of monotocous species, including cattle, yield multiple fetus births. Twin gestation and parturition can be classified into 2 categories: monozygotic and dizygotic. Monozygotic twins, also referred to as identical twins, result from the spontaneous cleavage of 1 fertilized oocyte during embryonic development. Evidence of monozygotic twinning in cattle is based on findings of concurrent embryos from reproductive tracts in which the ovaries contained only 1 corpus luteum (Lillie 1916) and the existence of conjoined twins (Johansson and Venge, 1951). Monozygotic twins are genetically and phenotypically the same and therefore can only be of the same sex. Dizygotic twins, also referred to as fraternal twins, result when 2 oocytes from 2 follicles become fertilized during the same estrous cycle. The distribution of same and mixed sex dizygotic twinsets reflects the expectation based on observed proportions of male vs. female embryos and independent determination of gender. In 1901, Weinberg’s differential method was created to determine the frequency of monozygotic twins in a population (Weinberg, 1901). In 1946, Gert Bonnier introduced a modification to Weinberg’s differential method that used the observed sex ratio in the sample population rather than Weinberg’s assumption of equal gender distribution (Bonnier, 1946).

Neither the Weinberg method nor the Bonnier method was employed in analysis of herd records in the current study because of potential skewing of same sex vs. mixed sex twin births due to the use of sexed semen. However, monozygotic twin births as a percentage of all twin births has been calculated previously for Holstein and Jersey breeds and their close relatives (Johansson and Venge, 1951; Meadows and Lush, 1957; Erb and Morrison, 1959; Johansson et al., 1974; Cady and Van Vleck, 1978; Nielen et al., 1989; Ryan and Boland, 1991), and estimates ranged from 4.47 to 13.7% for Holsteins and 3.1 to 4.69% for Jerseys. Proportions observed in this study of 3.5 and 9.1% for Jersey and Holstein breeds, respectively, were not significantly different from each other (P > 0.05) and are not directly comparable with the previous estimates from the literature, owing to a bias in the current study for greater availability of same sex (female-female) twinsets.

Placental vascular anastomosis is most known for enabling the freemartin condition. Freemartinism occurs in the female of mixed-sex multiple births, where the female’s reproductive tract develops abnormally. This abnormal development is caused by the presence of AMH, which is produced by the Sertoli cells of the male’s developing testis and affects the development of the Mullerian duct preventing normal formation of the uterus and oviducts in the female (Vigier et al., 1991). Freemartinism has been most notable in the bovine species, in both beef and dairy populations. Estimates of the percentage of freemartinism in heifers from mixed-sex twinsets ranges from 82.5 to 95.7% (Lillie, 1916; Zhang et al., 1994; Gregory et al., 1997). The observed frequency of leukochimeric twinsets in this study suggests a corresponding incidence of formation of anastomoses, and presumably freemartinism, had all twinsets been mixed sexed. The frequency of 80 to 85% is at the lower end of the previously reported range of freemartinism. None of the females from mixed-sex twin births in the current study were examined for freemartinism.

Summers et al. (1984) showed breed differences in the number of interspecific chimeric animals produced by embryo transfer, with Jersey and Brahman twinsets producing fewer chimeric twins vs. Holstein and Brahman twinsets. We hypothesized that a similar breed difference might be observed when comparing Jersey vs. Holstein twins. A statistically significant difference between breeds was not observed in the current study.

In the study by Summers et al. (1984), lymphocyte chimerism in Bos taurus taurus-Bos taurus indicus twins was found to not be stable. In most Friesian-Brahman twins, Friesian lymphocytes became increasingly dominant in each individual as the calves grew older. In 1 Jersey-Brahman pair, the percentage of Jersey lymphocytes in the Jersey calf remained stable, whereas in the Brahman calf, it increased from 15.9% at less than 1 mo of age to 43.7% at 12 mo of age (Summers et al., 1984). Data in the current study are insufficient to draw conclusions concerning changes in degree of leukochimerism with age. However, chimerism was clearly not transient, as all twins identified as chimeric at 2 d of age were identified as chimeric from 3 to 12 mo later.

The observation of several twinsets in both breeds having 1 twin with no chimeric SNP while its co-twin had at least 2 chimeric SNP may in some cases reflect the limits of the sensitivity of the assessment of nonchimerism. A twin whose leukocytes originate primarily from its co-twin will be readily identified as chimeric, while conversely a twin whose leukocytes originate primarily from itself may incorrectly be identified as nonchimeric. The experimental unit in the current study was twinset, and only 1 twin need be identified as chimeric to identify the twinset as chimeric.

The assessment of co-twin influence (Fig. 5) suggests 2 different relationships depending on the degree of leukochimerism. For twinsets with lesser leukochimerism (<45% co-twin influence for the most leukochimeric twin), there was not a significant (P > 0.05) relationship between co-twin influence in the twins. In contrast, for twinsets with greater leukochimerism, there was a significant (P < 0.001) inverse relationship between the co-twin influences of the twins. We speculate that the patterns seen in Fig. 5 reflect 2 different dynamics. The first is the extent of anastomosis or timing of formation of anastomoses relative to hematopoietic stem cell migration, and the second is a difference in timing of hematopoietic stem cell migration between members of a twinset. We hypothesize that where anastomoses are extensive and time of stem cell migration is similar, twins have approximately equal representation of self and co-twin. Where anastomoses are extensive and time of migration is dissimilar, 1 twin’s contribution predominates, leading to an inverse relationship in co-twin influence. Where anastomoses are less extensive, twins have high representation of self. The inverse relationship between percentage co-twin in members of a twinset, as described here, is equivalent to saying that proportionate representation of the 2 twins is similar for both members of a twinset. This has been observed in previous cytogenetic studies of male-female cattle pairs (Basrur and Kanagawa, 1969; Marcum, 1974; Plante et al., 1992). However, like Basrur and Kanagawa (1969), this study also observed a portion of twinsets (those with lesser leukochimerism) that did not follow this trend; these twins may have a limited degree of placental anastomosis, resulting in higher representation of self.

Assuming that genetic variation in the propensity for developing placental anastomoses exists, identifying the genes and polymorphisms responsible could facilitate selection against the formation of placental vascular anastomoses. Previous analysis of bovine twin pregnancy data suggests a dependency of postimplantation survival between bovine twins (Bennett et al., 1998), such that the loss of 1 fetus typically results in the loss of its co-twin, presumably due to the occurrence of vascular anastomoses and the resulting shared blood supply. We hypothesize that in the absence of placental vascular anastomoses, the loss of twin embryos during gestation could be independent, leading to higher pregnancy rates for twin gestations. Selection for reduced placental anastomoses could be complementary to selection for increased twining rate.





Be the first to comment.

Please log in to post a comment.
*Society members, certified professionals, and authors are permitted to comment.