ROCHESTER, New York—Control of colon cancer may be mediated in part by the patient’s immune system, suggesting that treatments that enhance this innate capability could aid in reducing mortality. This is the principal conclusion taken from work carried out by Alok Khorana, MD, and his colleagues at the James P. Wilmot Cancer Center, University of Rochester Medical Center, and presented at the American Association for Cancer Research 94th Annual Meeting (abstract 512).

The study involved immunocyto-chemical staining of paraffin-embedded stage II and III primary colon cancer tumors from 131 patients, seen at the University of Rochester from 1990 to 1995, and focused on vascular endothelial growth factor (VEGF), a protein that is overexpressed in a variety of cancers, often with an associated increase in severity of disease.

In addition to examining VEGF expression in the tumor itself, the investigators also measured its appearance in the surrounding tissue stroma, paying particular attention to its association with tumor-associated macrophages (TAMs). Sections were also stained for the antigens C68, a marker for macrophages, and CD3, a marker for T lymphocytes.

Immunostains were separately graded as the percentage of cells stained with moderate or greater intensity according to the following scale: negative, less than 2%; grade 1, 2% to 20%; grade 2, 20% to 50%; grade 3, greater than 50%. The degree of staining was further correlated with length of survival after resection of the primary tumor. Data were analyzed using a multivariate Cox proportional hazards model.

Study Results

Despite its frequent association with more severe disease, VEGF expression in the tumor itself was not significantly associated with survival. In contrast, there was a clear relationship between survival and VEGF expression in TAMs; 42% of patients showed this expression pattern, and median survival in this group was more than doubled, to 9.7 years vs 4.3 years in the VEGF-negative TAMs group (P = .0065).

Thus, rather than being associated with reduced survival, increased VEGF expression in one subset of tumor-associated cells actually predicted a more positive outcome. Confirmation that the stromal cells that expressed VEGF were indeed macrophages came from co-staining for the macrophage marker CD68; the great majority of VEGF-expressing cells were macrophages, with the remainder being fibroblasts. The VEGF-expressing TAMs comprised only a subset of the macrophages in the stroma, however, and in multivariate analysis, there was no association between survival and CD68 expression alone. It was accordingly inferred that the improved survival was a function of the presence of the VEGF-expressing TAMs subgroup, rather than of all macrophages.

One clue as to the possible mechanism mediating the link between increased survival and the presence of VEGF-expressing TAMs came from the examination of T-lymphocyte infiltration, as determined from the pattern of CD3 immunostaining. Only those tumors with VEGF-expressing TAMs showed significant infiltration of T cells (n = 10; P = .0003). Thus, it is possible that VEGF expression by TAMs is associated with mobilization of the T cells that are involved in a host immune response against the tumor.

"At present, we have only a correlation between VEGF-expressing TAMs and two other parameters, survival and T-cell infiltration, so that suggestions as to underlying mechanisms must still be regarded as speculative," he said. One possibility is that the macrophages are acting as antigen-presenting cells, and that T cells have been mobilized to the tumor in response to specific tumor antigens displayed by the macrophages. "While the correlation with VEGF expression seems clear, it remains a puzzle as to what role, if any, their VEGF expression plays in this process," he said. "It may be, in fact, that it is only acting as a surrogate marker of the activation of immune-related processes in given cells."

In this connection, he said, it has recently been reported that the transcription factor HIf-1α is essential for the activation of cellular inflammatory pathways (Cramer et al: Cell 112:645-657, 2003). "Since HIf-1a also upregulates VEGF expression, our correlations with increased patient survival may reflect the impact of overall activation of this subset of macrophages, not their VEGF expression as such," he said.

This could also explain why VEGF production by the tumor itself is not protective. However, he noted that VEGF is present in several isoforms, and the antibody used in the current study does not discriminate among them. It remains to be seen whether the macrophage-associated VEGF and the tumor forms are identical or different, he said.

"Obviously," Dr. Khorana said, "there is still a substantial amount of work to be done to understand these phenomena. Nonetheless, we believe that the most likely explanation for the improved survival associated with VEGF-expressing TAMs is an involvement of the host immune response in controlling the tumor. The implications are that further modulation of the immune response through therapeutic intervention may improve the prognosis for colon cancer." Some of this work has since been published (Khorana AA et al: Cancer 97:960-968, 2003).