The measurement of platelet-associated immuno–
globulins is one such method and these are discu–
ssed in detail later. An alternative test is the mea–
surement of reticulated platelets. These are imma–
ture platelets (Ault and Knowles, 1995), which
contain RNA and are analogous to reticulocytes
(reticulated erythrocytes). Reticulated platelets are
increased in conditions characterized by a high
platelet tumover such as ITP (Kienast andSchmitz,
1990) and decreased in conditions characterized
by bone marrow failure. For these reasons eleva–
tion of reticulated platelets preeectes the rise in pla–
telet counts during recovery from chemotherapy
(Romp et al, 1994).
1.3 Issues in measurement of reticulated plate–
lets: Reticulated platelets are commonly stained
with derivatives of thioflavin T such as thiazole
orange (TO) since it is excitable at the 488 nm la–
ser wavelength available in most flow cytometers
and also the dye perrneates the cell membranewit–
hout the need for permeabilization. TO provides
an approximately 3000-fold enhancement offluo–
rescence upon binding to RNA. The main diffi–
culty with the assay has been that the binding of
TO to RNA is not entirely specific and depending
on the protocol used TO could also bind to other
nucleic acids and even accumulate in dense granu–
les ofplatelets. Thus an ideal protocol should iden–
tify the cell type which is being studied, and use
appropriate controls during method development
and preferably during the assay itself to exclude
non-specific staining. The recent protocol byMa–
tic et al, 1998 meets these criteria. They identify
platelets by GPib staining and show that pretreat–
ment of sample with RNAse abalishes staining
whereas exposure to thrombin degranulates plate–
lets without a decrease in TO staining.
The other issue in the use of reticulated plate–
lets as a surragatemarker for ITP are the observa–
tions (Ault et al, 1992) in amurine ITPmodeland
in patients with elevated platelet associated im–
munoglobulins that the reticulated platelet count
correlates best with the state of thrombopoiesis
when the platelet counts are above 50,000/ul.
This samewhat limits the utility of the assay.
Platelet kinetic studies show that once the platelet
count dips to below 50,000/ul the platelet lifespan
is less than 3daysand by extrapolation is l-2 days
Klir1isk Kemi
i
Norden
4.
1999
at counts of 20,000/ul. At this point the survival
of normal platelets will approximate the duration
of time that a reticulated plateJet is identifiable by
its RNA content and reticulated plateletswill eon–
sritute a significant proportion of the circulating
platelets. They would thus become a major target
for destruction and thus result in a decrease in the
absolute levels of reticulated platelets.
2.1
Platelet antibodies:
Immunoglobulins or antibodies may bind to pla–
telets due to a variety of reasons. These include a)
autoantibodies against plateJet glycoproteins as in
chronic ITP, b) platelet specific antibodies, c) anti–
HLA antibodiesd) non-specific binding via plate–
let Fe receptors. I will briefly summarize the role
of flowcytometry in the measurement of these an–
tibodies.
2.2
Key issues in chronic ITP:
In chronic ITP the autoantibodies can bind to a
variety of glycoproteins though binding toGPIIb/
Illa and GPib/IX is the most common (Wadenvik
et al, 1998). The detection of platelet-associated
immunaglobulin by flowcytometry is a relatively
straightforward procedure. The problem is that
many patients with non-immune thrombocytope–
nias also show platelet surface immunoglobulins.
The assay thus has a low specificity since we are
unable to distinguish the specific binding of an–
tibodies
to
platelet glycoproteins such as GPIIb/
Illa from binding via Fe receptors on platelets
(Winiarski, 1998). To solvethis problem attempts
have been made tomeasure antibodies against the
specific platelet glycoproteins invalved such as
GPIIb/IIIa orGPib (McMillan et al, 1987). In the–
se assays lysates of patient platelets are used as
the source ofantigen and after suitable incubation
with patient plasma that presurnably contains an–
tibodies, the antigens are detected using known
anti-glycoprotein antibodies. While these assays
can identify the specific IgG antibodies against the
invalved glycoprotein in many cases, it is to be
realized that the assays willmiss those platelet gly–
coprotein antibodies not being screened (e.g. rarer
antibodies such as against GPIV), as weil as IgM
and IgA class antibodies, Fe receptor mediated
binding andHLAelass
l
antibodies. Thus, an intu–
itive approach may be to initially screen by FCM
93
