Risk Estimation for Balanced Translocation Carriers

1
Risk Estimation for Balanced Translocation
Carriers

• Carolyn Trunca, Ph.D.
• The Genetics Center
• Smithtown, NY

2
Introduction

• Carriers of balanced translocations are
  phenotypically normal, but they risk having
  abnormal children or miscarriages as a direct
  result of producing chromosomally unbalanced
  gametes.
• Carriers of balanced translocations are
  frequently given a general empiric risk estimate
  of between 10 and 20 of having a child with an
  unbalanced translocation. This implies that
  translocations can be treated as a homogeneous
  class.

3
Introduction

• However, we know from direct observations of
  meiosis in other organisms that the behavior of
  translocations, that is, the type and orientation
  of the multivalents at metaphase-1 and,
  therefore, the segregation at anaphase-1, depends
  to a great extent on the structural
  characteristics of the translocation.
• We, also know, that the likelihood that an
  abnormal segregation will result in the birth of
  child with an unbalanced translocation depends on
  the viability of
• the resulting partial trisomies and partial
  monosomies.

4
Purpose

• The purpose of this project was
• to collect cytogenetic and reproductive data from
  a large number of families in which a
  translocation is segregating
• to use that data to identify those factors that
  are associated with poor reproductive outcome in
  translocation carriers, and
• 3) to use that information to develop a
  method to estimate both the risk a carrier of a
  specific translocation has of having a child with
  an unbalanced translocation and the risk of
  having a miscarriage.

5
Background

• Meiosis in translocation carriers proceeds
  according to the following principles.
• At pachytene, each segment of the rearranged
  chromosomes will pair with the homologous segment
  of the normal chromosomes forming a pachytene
  cross.
• 2. At diplotene, chiasmata will appear where
  crossing over has occurred. The number of
  chiasmata is positively correlated to the length
  of the arm of the pachytene cross. It is
  possible that a very short arm may have no
  chiasmata.

6
Segregation of a Reciprocal Translocation
7
Background

• 3. At diakinesis and first metaphase, the four
  chromosomes are associated as either
• a ring (having at least one chiasmata in each
  arm)
• a chain (having at least one chiasmata in 3 of
  the 4 arms)
• or be separated into two bivalents, a univalent
  and a trivalent, two univalents and a bivalent or
  even four univalents (each possibility a result
  of the failure of chisamata formation in 2 or
  more arms).

8
Segregation of a Reciprocal Translocation
9
Segregation of Translocations That Are
Predisposed to Form Type I or Type II Chains
10
Background

• 4. At anaphase, there are three ways that the
  four chromosomes can concordantly orient so that
  two go to each pole of the spindle
• Alternate alternate centromeres in a ring or
  chain pass to the same pole so that all gametes
  are genetically balanced, 50 will contain the
  two normal chromosomes and 50 the two
  translocated chromosomes

11
Background

• Adjacent-1 adjacent non-homologous centromeres
  pass to the same pole, so that all gametes are
  genetically unbalanced
• 2) Adjacent-2 adjacent homologous centromeres
  pass to the same pole, so that all gametes are
  genetically unbalanced

12
Segregation of a Reciprocal Translocation
13
Background

• 5. At anaphase, two bivalents will segregate
  independently, so that 50 of the gametes will be
  genetically balanced and 50 unbalanced.
• 6. At anaphase, if a univalent and a trivalent
  have occurred, a 31 disjunctions may occur and
  then all gametes will be unbalanced.
• 7. At anaphase, if there is a discordant
  orientation of a ring, a 31 disjunction can
  occur. This occurs when two centromeres on
  opposite sides of the ring pass to opposite poles
  , while the two intermediate centromeres pass to
  the same pole.

14
Research Design

• DATA COLLECTION
• Data were collected from 824 families in which a
  balanced autosomal translocation with
  identifiable breakpoints was segregating.
• Families with Robertsonian translocations,
  translocations involving sex chromosomes, and de
  novo translocations were not included.
• Most of the families were obtained from the
  literature, but personally studied families, and
  unpublished families contributed by other
  investigators were also included.

15
Cytogenetic Data

• Calculated for this study for both chromosomes in
  a translocation were
• the lengths of the short arm and the long arm,
• the distance from the breakpoint to the end of
  the arm in which the break occurred (terminal
  distance),
• the distance from the breakpoint to the end of
  the other arm,
• the distance from the breakpoint to the
  centromere (interstitial distance).

16
Cytogenetic Data

• These lengths were were determined by
  measurement, in arbitrary units, of the published
  diagrams representing chromosomes at the 450 band
  level.
• As an example, for a translocation between
  chromosomes 6 and 21 with breaks at 6q22 and
  21q22, the following values were recorded.
• Chromosome 6 p79.4, q125.8, n41.5, c163.7,
  i84.3
• Chromosome 21 p20.2, q46.00, n9.05, c57.15,
  i36.96

17
Research Design

• DATA COLLECTED ON THE FAMILY
• Method of Ascertainment
• 1. A child with an unbalanced translocation
• 2. History of Multiple Miscarriages or
  Infertility
• 3. Chance

18
Research Design

• DATA COLLECTED ON INDIVIDUALS
• Proband, Excluded, Included?
• Live-born or Fetal Death
• Gender
• Phenotype
• Karyotype
• Number of Chromosomes Type of
  Segregation

19
Research Design

• CORRECTING FOR ASCERTAINMENT BIAS
• The proband or probands are excluded from the
  database.
• Individuals in the family that bring another
  sibship into the database are excluded.

20
Typical pedigree
21
Methods

• These data were first analyzed to determine if
  structural factors influence the type of
  segregation that occurs during meiosis in human
  translocation carriers.
• We approached this problem by testing six
  hypotheses based on observations in other
  organisms where direct analysis of meiosis is
  possible.

22
Hypothesis

• 1. Whatever the type of segregation that occurs
  at the first meiotic anaphase, normal gametes and
  gametes carrying the two translocated chromosomes
  occur with equal frequency.

23
Corrected Number of Normal vs. Carrier Offspring
of Translocation Carriers

• Normal Carrier
• 777(48.2) 835 (51.8)
• X2 2.02 p
  0.15

24
Hypothesis

• 2. An interstitial chiasma forces the linked
  homologous centromeres to co-orient so adjacent-2
  segregation are rare.

25
Translocation With Long Interstitial Segments

• Adjacent-2 segregation are rare because
  homologous centromeres co-orient.

26
Frequency of Adjacent Segregations At least One
Long Interstitial Segment

• Adjacent-1 Adjacent-2 ( Adjacent-2)
• 347 3
  (0.8)

27
Hypothesis

• 3. In the absence of interstitial chiasmata any
  two adjacent centromeres can co-orient,
  therefore, among adjacent orientations both
  adjacent-1 and adjacent-2 segregations occur.

28
Translocation With Two Short Interstitial Segments

• In the absence of interstitial chiasmata any two
  adjacent centromeres can co-orient so among
  adjacent segregations, both adjacent-1 and
  adjacent-2 occur.

29
Frequency of Adjacent Segregations Two Short
Interstitial Segments

• Adjacent-1 Adjacent-2 ( Adjacent-2)
• 218 25 (10.3)

30
Hypothesis

• 4. When Type 1 chains form because chiasmata are
  absent from one or the other of the interchanged
  arms, the stable adjacent orientation that occurs
  is adjacent-1.

31
Translocation Predisposed to Forming a Type I
Chain and Undergoing Adjacent-1 Segregation

• Failure to Undergo Recombination Occurs in Short
  Exchanged Arm of Pachytene Cross

32
Segregation of Translocations That Are
Predisposed to Form Type I or Type II Chains
33
Frequencies of Types of Adjacent Segregations
Among Translocations That Tend To Form Type I
Adjacent-1 Adjacent-2 ( Adjacent-2)
531 4
(0.8)
34
Hypothesis

• 5. When Type 2 chains form because chiasmata are
  absent from one or the other of the
  non-interchanged arms, the stable adjacent
  orientation that occurs is adjacent-2.

35
Translocation Predisposed to Forming a Type II
Chain and Undergoing Adjacent-2 Segregation

• Failure to Undergo Recombination Occurs in Short
  Non-exchanged Arm of Pachytene Cross

36
Segregation of Translocations That Are
Predisposed to Form Type I or Type II Chains
37
Frequencies of Types of Adjacent Segregations
Among Translocations That Tend To Form Type
II Adjacent-1 Adjacent-2 (
Adjacent-2) 11
26 (70.3) Represents
76.5 of all adjacent-2 segregations found in
the total sample.
38
Hypothesis

• 6. If the pachytene cross is highly asymmetrical
  with two short arms and short interstitial
  segments, the possibility of a 31 disjunction at
  anaphase is increased. This occurs as a result
  of either
• 1) a discordant orientation of a chain when
  crossing-over fails to occur in one of the short
  arms or
• 2) the formation of a trivalent and a univalent
  when crossing-over fails to occur in both of the
  short arms. In this case the univalent will pass
  at random to either pole.

39
Asymmetrical Pachytene Cross Predisposes
Translocation to Undergo a 31 Disjunction

• One Exchanged Arm and One Non-exchanged Arm of
  the Translocation Is Very Short

40
Frequency of 31 Disjunctions Highly Asymmetric
Pachytene Cross vs. Symmetric Pachytene Cross

• 31
  Adjacent
• Disjunctions Segregations
  ( 31)
• Highly Asymmetric 63
  25 (71.6)
• Pachytene Cross
• Relatively Symmetric 156 626
  (19.9)Pachytene Cross

41
Question

• The confirmation of these hypotheses is
  significant in that it indicates that the
  structural characteristics of a reciprocal
  translocation are important in determining how
  meiosis will proceed. Any attempt at developing
  risk estimates must take structural factors into
  consideration.
• Can we make predictions about which
  translocations are likely to pose a low risk for
  having a child with an unbalanced translocation
  and which confer a high risk?

42
Low Risk Translocation

• Large Pieces Exchanged Predisposes Translocation
  to Undergo Alternate Segregation

43
High Risk Translocation

• Very Small Pieces Exchanged Predisposes
  Translocation to Undergo Independent Assortment

44
Risks of Abnormal Offspring and Fetal Deaths for
Carriers of Two Structurally Different
Translocations

• Abnormal Child Fetal
  Death
• Small Pieces Exchanged 25.1 2.2 (375)
  23.8 (558)
• Prediction HIGH RISK
• Large Pieces Exchanged 1.6 1.1 (127)
  30.8 (237)
• Short Interstitial Segments
• Prediction LOW RISK

45
Question

• There are likely to be selection differences
  between chromosomally unbalanced eggs and sperm.
• Are there differences in reproductive risk
  between female and male translocation carriers?

46
Risk of Abnormal Offspring for Male and Female
Carriers

• Male Female
• Underestimate 6.6 (966)
  11.9 (1539)
• Overestimate 14.9 (1060) 18.6
  (1665)

47
Question

• How a family is ascertained may give some
  indication of the likelihood that, if an abnormal
  segregation occurs, it will result in the birth
  of a child with an unbalanced translocation.
• Are there differences in risk for families in
  the different ascertainment groups?

48
Risk of Abnormal Offspring by Ascertainment
Groups

• Abnormal
  Subfertile Fortuitous
• Underestimate 11.9 (2078) 1.3 (151)
  0.4 (547)
• Overestimate 20.3 (2297) 5.7 (158)
  2.7 (560)

49
Question

• The structural characteristics of a translocation
  depend on which chromosomes are involved and
  where the breakpoints are located within the
  chromosomes. Risk depends on the structural
  characteristics and the viability of the abnormal
  segregants. Can we determine which chromosomes
  and which breakpoint locations increase risk?
• Since we could show that there are obvious
  differences in risk between translocations
  ascertained either through an abnormal child,
  through a history of multiple miscarriages or
  infertility, or by chance, all our analyses were
  performed on the sample as a whole, and on each
  ascertainment group separately.

50
Analysis

• Null Hypothesis Chromosomes involved in
  translocations occur at random.
• Statistical analysis of the entire data set
  indicated that chromosomes are not involved in
  translocations at random and the difference is
  highly significant.
• To investigate this observation further, the data
  from each of the three data sets were analyzed
  separately.

51
Chromosome Involvement in Translocations by
Ascertainment Groups Abnormal
Subfertile Fortuitous
(N557) (N89)
(N123) 9, 11, 13,
18, 21, 22 22 None
Overrepresented 1, 2, 3, 6,
7, 19 None None
Underrepresented Level of significance 0.002 or comparisons
52
Conclusions

• These results indicate that breakage and,
  therefore, chromosome involvement in
  translocations occurs randomly.
• The significant differences from randomness
  observed are the result of differences in risk.
• Carriers of translocations involving chromosomes
  9, 11, 13,18, 21, and 22 are more likely to have
  an abnormal child, and are, therefore, more
  likely to be included in the database.
• The reverse is true for translocations involving
  chromosomes 1, 2, 3, 6, 7, and 19.

53
Analysis

• Null Hypothesis Breaks are distributed at
  random within a given chromosome, that is, the
  proportion of breaks within a particular band
  equals the length of the band divided by the
  total length of the chromosome. This implies a
  uniform probability distribution.
• The combined sample gives a poor fit to random
  allocation of breaks for essentially every
  chromosome. This reflects the fact that the
  breakpoints distribution is non-random in the
  chromosomes from translocations ascertained
  through an abnormal child.

54
Non-random Distribution of Breaksin Chromosome 1
when Translocation Was Ascertained Through an
Abnormal Child
55
Random Distribution of Breaksin Chromosome 1
when Translocation Was Fortuitously Ascertained
56
Conclusions

• This observation indicates that
• 1. in an unselected population, chromosome
  breakage occurs randomly along the length of
  chromosomes
• 2. the distribution of breakpoints along the
  chromosomes differs significantly between
  ascertainment groups and
• 3. the non-random distribution of breakpoint
  locations observed in chromosomes from
  translocations ascertained through an abnormal
  child is the result of the increased risk that
  translocations with terminal breaks confer.

57
Analysis

• Since there is a clustering of breakpoints toward
  the ends of the chromosome arms in the abnormal
  group, the terminal distance, that is, the
  distance from the breakpoint to the terminal end
  of the chromosome arm, was measured as
  quantitative data.
• The mean terminal distances for all chromosomes
  in all groups were analyzed to determine if there
  were significant differences in the mean terminal
  distance between chromosomes and between
  ascertainment groups.

58
Analysis
59
Conclusions

• In the fortuitous group, the mean terminal
  distance increases as the size of the chromosome
  increases. This is as expected since the
  breakpoints are randomly distributed for this
  group.
• In the abnormal group, the mean terminal distance
  is the same for each chromosome and, on average,
  shorter than the mean terminal distance observed
  in the other groups.
• This suggests that the terminal distance is
  another important factor in determining
  reproductive risk.

60
Final Conclusions

• The variables that we have shown to be predictors
  of the risks for having a child with an
  unbalanced translocation or a miscarriage are
• 1. the chromosomes involved in the
  translocation
• 2. the gender of the carrier
• 3. the way the family was ascertained and
• 4. the length of the terminal distances.

61
Risk Estimation

• Logistic regression, like linear regression
  produces prediction equations. It is used to
  determine whether specific factors are related to
  the presence of some characteristic, for example,
  whether having chromosome 7 in a translocation is
  predictive of having an abnormal child.
• Unlike linear regression which can be solved
  explicitly, that is, there is a formula for it,
  logistic regression equations are solved
  iteratively. A trial equation is fitted and
  tweaked over and over to improve the fit.
  Iterations stop when the improvement from one
  step to the next is negligible.

62
Risk Estimation

• The response variable that characterizes logistic
  regression is an indicator of the presence or
  absence of a characteristic, that is, a (yes/no)
  variable.
• A logistic regression equation does not directly
  predict the probability of an occurrence. The
  output of a logistic regression equation is a
  constant and a coefficient for each predictor
  variable expressed as a log odds. Each
  coefficient represents the change in the response
  (e.g., the increase or decrease in risk) per unit
  change in the predictor (e.g., terminal length).
  While a log odds output at first does not seem
  very helpful, it is possible to transform a log
  odds to a probability.

63
Risk Estimation

• Logistic Regression Equation for Estimating the
  Risk of a Child with an Unbalanced Translocation
• Female carrier of t(67)(q23q11) with multiple
  miscarriages
• Coefficients
• Constant (-5.412), Chrom. 6 (0)
  Chrom. 7 (-0.616)
• Term. 6 (-0.6237) Term. 22 (-0.2727)
• Female (1.12) Asc. (1.76) -3.96
• Log Odds -3.96 Odds 0.019
• RISK () (Odds/1 Odds) x100 0.019/1.019
  x100 1.8

64
Acknowledgements

• Christa Ugrinsky
• David Wiener
• Amy Kaplan
• Liz Meller
• Martha Vlasits
• John Milazzo
• Nancy Mendel