The LAD known as LAD-1 is caused by reduced β2 integrin expression on leukocytes. LAD-1 has been recognized as a neutrophil disorder because of recurrent life-threatening infections principally due to the lack of the phagocytic receptor Mac-1 (9
). We describe patient JT, who has all the clinical features of LAD-1 but has expression of the β2 integrins at levels similar to those of LAD-1 heterozygotes, who have ~50% of normal expression but are clinically unaffected. Two novel mutations, S138P and G273R, were identified in the patient's CD18 alleles. We have shown that S138P can support expression of LFA-1 and Mac-1 on K562 transfectants but that the expressed integrin does not function. The G273R mutation supports neither expression nor function of LFA-1 or Mac-1. It is therefore likely that the CD11/CD18 antigens expressed on the patient's leukocytes contain the S138P mutation in the CD18 subunit and that the G273R mutation is responsible for the LAD-1 heterozygote-like levels of expression.
Neither the patient's neutrophils nor K562 transfectants expressing Mac-1(S138P) adhere to Mac-1 ligands fibrinogen and iC3b. However, the patient's T cells have limited CD11a/CD18–mediated adhesive properties. Whereas they do not adhere to ICAM-1 under any of the activation protocols tested, they can adhere to ICAM-2 and ICAM-3 when activated extracellularly with Mg2+
/EGTA and Mn2+
. This result provides evidence that LFA-1 recognizes each ligand in a unique way and suggests the possibility that there could be a structural defect in the patient's LFA-1 that prevents interaction with ICAM-1 but allows some interaction with ICAM-2 and ICAM-3. There is previous evidence that LFA-1 interacts distinctively with its individual ligands. For example, CD11a MAB MEM-83 induced LFA-1 to selectively bind ICAM-1 but not to ICAM-3 (33
), and, conversely, two other CD11a MABs block binding of LFA-1 to ICAM-3 but not to ICAM-1 (34
). The binding site for LFA-1 on ICAM-1 is located within the first two domains (21
); however, it appears that domain 1 of ICAM-3 is sufficient for LFA-1 binding (36
). Electron microscopic evidence shows that for ICAM-3, domain 2 is less structurally dependent on domain 1 than for ICAM-1(35
). Another difference is that ICAM-1 is present on the membrane as a dimer (39
), whereas at least ICAM-2 appears to be a monomer (40
). Finally, the fact that murine LFA-1 can bind human ICAM-2 and ICAM-3, but not ICAM-1, implies that the first two ligands might have structural features in common not shared with ICAM-1 (41
). Taken together, the implication of these observations is that the interaction of LFA-1 with ICAM-1 has special features distinct from that interaction with ICAM-2 and ICAM-3.
The patient's EBV-transformed B cells, and K562 transfectants expressing LFA-1(S138P), also failed to adhere to ICAM-1, but, unlike the patient's T cells, they cannot be promoted to adhere to ICAM-2 and ICAM-3 by divalent cation manipulation. This discrepancy may be due to the inability of the EBV-transformed B lymphocytes and the transfectants to respond to the full range of integrin-inducing signals. For example, adhesion to ICAM-2 or ICAM-3 of EBV-transformed B cells from a normal individual, and K562 cells expressing wild-type LFA-1, can only be consistently demonstrated with Mg2+
/EGTA in addition to an activating MAB such as KIM 185. In comparison, either Mg2+
/EGTA or KIM 185 is sufficient to activate ICAM-1 adhesion (Table and ref. 26
). The patient's T cells had a completely normal ability to bind to fibronectin via the α4β1 (CD49d/CD29) and α5β1 (CD49e/CD29) integrins. As with other patients with LAD-1, the ability of leukocytes to bind to other integrins such as the fibronectin receptors makes possible lymphocyte responses, and, for this patient, the additional ability of recognizing ICAM-2 and ICAM-3 may also have contributed to his survival into the midteen period.
The mutations found in the β2 subunit of the patient have not been described previously. Both mutations lie in the putative I domain of the integrin β2 subunit. A model of this region of the β2-subunit I domain has two extra β strands between β strand D and helix 5 (42
) when compared with a solved α-subunit I-domain structure (43
). The G273R mutation that leads to nonexpression of integrin is located in the loop between β strand D and the first extra β strand. No function has yet been assigned to this loop, but it is of interest that G273R mutation affects heterodimer formation.
The S138P mutation is novel in that it gives rise to expression of αβ heterodimer, but these integrins are without function. The S138P mutation is also novel in that it is the first reported natural mutation within the metal ion–dependent adhesion site (MIDAS) motif, which lies between the β strand A and helix 1 of the I-domain model. The divalent cation binding MIDAS-like (DXSXS) motif was initially described in the I domain of the α subunit as the key Mg2+
binding site of the integrin on which ligand binding totally depends (43
). It is also appreciated that β subunits contain a MIDAS motif that binds cations and is involved in ligand binding (42
). The patient's mutated S138P represents the second Ser in the DLSYS of the β2 MIDAS motif. The importance of the conserved residues in DLSYS motif of β2 was studied by mutagenesis in which the D134, S136, and S138 were replaced with Ala in three separate constructs (46
). Whereas the D134A and S136A mutants were fully capable of forming LFA-1 and Mac-1, the resultant heterodimers were incapable of binding ICAM-1 and iC3b, respectively. Whether the mutants LFA-1(D134A) and LFA-1(S136A) can bind ICAM-2 and ICAM-3 was not addressed. The S138A mutant, in this same work, was reported to be defective in forming LFA-1 and Mac-1 (46
). However, the same mutation was reported to be competent in forming Mac-1 by another group. The Mac-1(S138A) was not adherent to iC3b, denatured ovalbumin, and fibrinogen, but its binding to the neutrophil inhibitory factor was not impaired (47
Another indicator of the status of integrin function is the expression of integrin “activation reporter” epitopes such as detected by MAB 24 for β2 integrins (7
). Expression of this epitope correlates with the increased ability of LFA-1 to bind soluble ICAM-1, which is taken as a measure of high-affinity integrin (7
). Expression also correlates with an interdomain movement involving the α subunit I domain that occurs upon binding of Mg2+
). The lack of 24 epitope expression by the β2 integrins on the patient's neutrophils and T cells confirmed that these integrins were not undergoing activation and may be deficient in tertiary structural changes necessary for full integrin function. Although Mg2+
/EGTA caused neither 24 epitope expression nor adhesion to ICAM-1, it did promote adhesion to ICAM-2 and ICAM-3. The speculation here is that 24 epitope expression may mark an LFA-1 conformation that is specifically required for ICAM-1 binding. In this regard, it is interesting that the mutagenesis of the β3-subunit residue S123A, which corresponds to S138 of β2, is associated with a lack of ability of the integrin αIIbβ3 to undergo conformational change (49
In summary, we describe a patient expressing adequate levels of β2 integrin but with failure of β2-integrin function that has led to disease that is clinically indistinguishable from a moderately severe form of LAD-1. This patient differs from another recently described variant LAD-1 that has normal CD18 alleles but may have a signaling defect (50
). Patient JT is, however, similar to a Glanzmann's thrombasthenia patient who has platelets with heterozygote-like expression of αIIbβ3. On one allele, this patient has a mutation giving rise to a truncated β3 subunit, leading to faulty signaling and loss of function (51
). Because patients with defects in integrin function may not be easily recognized by the usual characteristic of absent or low expression, it is presently unknown how common this variant type of lesion might be.