Telomere dynamics of GPI^– and GPI⁺ granulocytes from patients with PNH. (A) Immunomagnetic separation of GPI^– and GPI⁺ granulocytes from a patient with PNH (left panel). By staining with anti-CD59, one sees the efficient separation of GPI^– (middle panel) and GPI⁺ (right panel) granulocytes. This technique enabled us to recover both the GPI^– and the GPI⁺ granulocytes with a purity higher than 90%, as assessed by flow cytometry after staining with anti-CD59. (B) Southern blot analysis of the telomere length of GPI^– and GPI⁺ granulocytes from 3 patients with PNH. Each patient is indicated by a capital letter (Table 1). Ten micrograms of high molecular weight genomic DNA, digested to completion with BamHI, was resolved on 1% agarose gel by field-inversion gel electrophoresis. After blotting, the filter was hybridized with TelBam8, a probe that is unique for the subtelomeric region of the long arm of chromosome 7.²³  The length of the BamHI telomeric fragment was calculated from the densitometric profile. We have previously shown that by this method, in a side-by-side comparison, differences greater than 320 bp are significant.²⁰  (C) Telomere length in GPI⁺ and in GPI^– granulocytes from individual patients. The differences in telomere length between paired DNA samples of GPI^– and GPI⁺ granulocytes from each PNH patient are shown by individual straight lines. Each patient is indicated by a capital letter (Table 1). (D) Scattergrams of the TRF^O–E (difference between the observed telomere length, TRF^O, and the expected-for-age telomere length, TRF^E) in GPI^– and GPI⁺ granulocytes from PNH patients, and in 40 age-matched healthy individuals (controls). The horizontal lines across each set of data points are the median values.