Engraftment monitoring: part I

Following transplant of hematopoeitic stem cells, one must carefully follow up the patient for months. This is because the aim of a bone marrow transplant is to reconstitute the immune system, or - to be more precise - to borrow some immunity from a donor. Which means, eventually, that at some point of time, most of immune system (i.e. the white blood cells) of the recipient has to be replaced entirely by the donor's. In essence, the recipient will become a genetic chimera, carrying a distinct cell population with the donor's genome.

Now, how do you know when that happens? White blood cells look all alike under the microscope, and there is no telling whether it belongs to the patient or donor just by looking at them. Which brings us to they key question of identity, i.e. how am I any different from you?

If the donor and recipient are of opposite gender, a simple karyotype from a sample of blood cells would easily identify which one is from whom. The Y chromosome will stand out. However:

1. If you have ever done cytogenetics, you know the mind-numbing labor that goes into karyotyoing

2. If multiple donors of different genders have donated stem cells to a single recipient (as is often the case), the method falls on it face.

So let's move on to genetic differences between two individuals. As members of the same species, we are more or less similar, except for a few things. Short tandem repeats (STR) are repetitions of a short nucleotide core (usually less than 9 nucleotides), which vary in repeat numbers between people. For example the donor might have:

GTCGTCGTCGTC

is a 4-repeat of a 3-sequence. Again, the recipient could have:

GTCGTCGTCGTCGTCGTCGTC

is a 7-repeat of the same. Which is to say that if you perform a polymerase chain reaction for this particular STR, the band for the donor and the recipient would show up at different places.

For small differences like a few nucleotides, gel electrophoresis is too low res. On capillary electrophoresis, the same would show up as two peaks.

Capillary electrophoresis of STR PCR; the exact height and area of the peaks are subject to PCR conditions, the loci being chosen and the relative difference in repeats between the two

Remember: PCR is competitive; shorter segments will eat up reagent nucleotides faster than longer ones (simply because the longer repeats need more time to copy). In this case, the difference 7-4=3 is not much; but if you pick a VNTR (variable number of tandem repeats) which are much more variable in repeats than STR, the difference can show up and skew the analysis in favor of the shorter allele.

Like every other gene, STRs are also inherited in two alleles: one maternal and one paternal. A person might have

• homozygous at an STR locus, i.e. both alleles might have the same number of repeats
• heterozygous, i.e. one allele has a 7-repeat, another a 3-repeat (i.e. the aforementioned scenario)

In an ideal scenario, the donor and the recipient have completely different STR alleles: maybe the donor has a 7-repeat and a 4-repeat, and the recipient has a 10-repeat and a 15-repeat. Which makes things really simple to analyse:

The proportion of donor cells in this case is simply the proportion of donor alleles (a & b) in the post transplant mix (green), i.e. 

The height/ area of the peaks reflect the relative proportion of the alleles in the mix (remember: all are competing for the same primer!)

(a + b)/ (a + b + c + d)

This simplest scenario is of course, quite idealistic. One won't often find such an STR between two individuals; and even if one does, it is not wise to assess engraftment by only one STR marker. Consider this particular situation where the exact same allele pair is carried by both donor and recipient.

This is a non-informative STR marker - because the donor and recipient carry the same allele pair

In other situations, there might be one common allele between the two.

The allele at extreme right is ahred between the recipient; the one at extreme left belongs to donor only. Thus, the proportion of the left allele in the final sample is the amount of chimerism

This is akin to the schema:

Here, the amount of chimerism is b / (b + d), i.e. pretend that 'a' is not there.