Clinicopathologic analysis of macrophage infiltrates in breast carcinoma

We compared macrophage density, assessed by enumeration of peritumoral mononuclear cell immunoreactivity for HAM 56, to clinicopathologic features and to immunostaining for two "invasion-associated" proteases (Cathepsin D and Urokinase plasminogen activator) in 80 breast carcinomas. Diffuse (2+) infiltrates of HAM 56- positive mononuclear cells were present in 27 cases (34%) and 43 (54%) exhibited focal (1+) infiltrates. Presence of 2+ macrophage infiltrates correlated significantly with poor differentiation. None of the seven well-differentiated cases exhibited 2+ infiltrates, whereas 9/43 (21%) moderately differentiated and 18/30 (60%) poorly differentiated tumors were diffusely infiltrated (p = .001). Wide-spread macrophage infiltrates were also more frequent in cases with advanced stage (23% of node negative vs 40% of node positive cases, p = NS). Forty-four percent of the cases with diffuse macrophage infiltrates were cathepsin D positive (i.e. in host derived cells) vs only 18% with focal macrophage infiltrates (p = .002). A similar relationship was observed between staining for HAM 56 and urokinase-type plasminogen activator (p = .02). Disease recurrences (50 months median follow-up) were more frequent in patients with 2+ (17/27, 63%) as opposed to 0+ (1/10, 10%) macrophage infiltrates (p = .01). We conclude that the density of stromal macrophage infiltrates is associated with clinical aggressiveness in breast carcinomas. Further, this relationship may reflect contribution of host derived macrophages to invasion and metastasis through elaboration of proteases which putatively mediate degradation and remodeling of extracellular matrix.