Get documents about "The Pathway to Biomarker Discovery Carbonic Anhydrase IX and
Document Sample
Biology Behind
The Pathway to Biomarker Discovery: Carbonic Anhydrase IX and
the Prediction of Immune Responsiveness
55Commentary on Atkins M et al., p. 3714
Monica C. Panelli, Ena Wang, and Francesco M. Marincola
The relevance of identifying predictors of response and/or
Relevance of the Question
toxicity to IL-2 therapy, however, goes beyond the practical
The level of carbonic anhydrase IX-G250MN (CAIX) aspects discussed above as this cytokine may be a central
expression in renal cell carcinoma (RCC) tissue samples may modulator not only of melanoma and renal cell cancer
serve as the first useful biomarker predictive of a response to regression, but it might be implicated in other important
interleukin 2 (IL-2) therapy (1). IL-2 is the only biological clinical phenomena including transplant rejection. Recent work
agent approved by the U.S. Food and Drug administration for has shown that the administration of IL-2 receptor antagonists
the treatment of stage IV RCC because it can induce in a small decreases the incidence of kidney allograft rejection (6). This
percentage of patients’ durable complete responses (2). The use clinical observation seems to link the immune biology of allo-
of IL-2 is limited by severe toxicity and cost; therefore, the transplantation with that of immune-mediated cancer rejection
therapeutic value of the drug would be enhanced by the ability and goes in line with the recent observation that the genetic
to improve the risk/benefit ratio of its administration by profile of a melanoma metastasis responding to IL-2 therapy
predicting which patients may or may not be responsive. (7) overlapped with that of kidney tissue undergoing acute
Recent efforts have been focused in identifying biomarkers rejection (8). Therefore, a deeper understanding of the biology
predictive of response to therapy that could be used for of IL-2 when given to humans in vivo may have broader
patients’ selection. The level of CAIX expression in RCC tissue implication than its practical application in the decision
samples may serve as the first useful predictive marker (1) making process during patient enrollment (9).
because the association of most other factors with clinical
response is not universally accepted (2) and for several of them The Biology Behind
it is treatment related (i.e., toxicities) and therefore cannot be
used for patient selection (2, 3). In particular, when the CAIX- The mechanism(s) at the basis of clinical regressions in
based model was refined by stratifying patients according to response to systemic IL-2 administration remain(s) elusive. In
other favorable or unfavorable pathologic features, the fact, we observed that the in vivo activity of this biological
prediction was almost faultless (see Table 4 in Atkins et al.; response modifier seems to differ from its predicted effect on
ref. 1). Whereas the results of this study should encourage the T-cell stimulation. IL-2, instead, mediates an activation of
choice of systemic IL-2 therapy for patients with CAIX innate effector mechanism through the downstream release of
expressing tumors, we agree with Atkins et al. (1) that the an array of cytokines by circulating and/or intratumoral
overlap in response rates observed among patients with immune cells (7, 9). Transcriptional analysis of melanoma
different levels of CAIX expression should not justify its use metastases in a relatively small cohort of patients undergoing
as an absolute criterion for patients’ exclusion from treatment combined antigen-specific immunization and systemic IL-2
pending further independent and prospective validation. In administration suggested that immune responsive lesions
addition, relationships have been described between toxicity are immunologically different even before treatment from
and effectiveness of systemic IL-2 administration suggesting a immune-resistant lesions because they express immunologic
common effector pathway (4, 5). Therefore, it would be signatures consistent with chronic inflammation (10). IL-2
interesting to know in the future whether CAIX is associated may act through the downstream release of soluble factors
with some of the IL-2 related toxicities or it is an independent including cytokines and chemokines (9, 11) that may turn
predictor of responsiveness. Obviously, the ability to segregate a chronic inflammatory process into an acute one through
the mechanism(s) responsible for these two biological effects activation of innate immune effector functions (7). In
of IL-2 may suggest ways of administration as single agent or in particular, this acute inflammation may act as a costimulatory
combination therapy to increase the rate of responses whereas signal for tumor antigen-specific T cells that have reached the
decreasing toxicity. tumor site and are exposed to antigen stimulation by
interactions with tumor cells or immature mononuclear
phagocytes (12, 13). However, confirmed biomarkers of
immune responsiveness are not available in the context of
Authors’ Affiliation: Immunogenetics, Section, Department of Transfusion either melanoma or RCC.
Medicine, Clinical Center, NIH, Bethesda, Maryland
Received 3/2/05; accepted 3/22/05.
Requests for reprints: Francesco M. Marincola, Department of Transfusion How Does CAIX Fit the Picture?
Medicine, Clinical Center, Building 10, Room 1C-711, 10 Center Drive MSC 1502,
Bethesda, MD 20892-1502. Phone: 301-496-9702; Fax: 301-594-1981 E-mail:; In this issue of Clinical Cancer Research, Atkins et al. (1)
FMarincola@ mail.cc.nih.gov. expand on previous observations (2, 14) suggesting that CAIX
www.aacrjournals.org 3601 Clin Cancer Res 2005;11(10) May 15, 2005
Biology Behind
Fig. 1. Eisen’s clustering based on 345
genes significantly (t test P 2 < 0.001)
up-regulated in RCC compared with normal
kidneys. A node containing CAIX is
enlarged to outline genes coordinately
expressed. The relationship among
experimental samples (all primary RCC
samples) is shown by the clustering tree on
top of the enlarged image. For
methodologic details, please refer to
ref. 15.
is a favorable prognostic marker for RCC and its constitutive CAIX is an isoenzyme of the large family of carbonic
expression in primary or metastatic lesions may be a predictor anhydrases that may have redundant function. For instance,
of immune responsiveness to IL-2 therapy (14). We recently we noted that metastatic melanomas express specifically
noted by gene profiling of tumors of various histology that the carbonic anhydrase III (muscle specific) and XIV which may
expression of CAIX is mostly confined to RCC (15). However, have redundant function and compensate for the lack of CAIX
CAIX was sporadically expressed by other epithelial tumors but (15, 16).
never by normal kidneys suggesting that its transcriptional Atkins et al. (1) suggested that CAIX may maintain a
activation, contrary to that of CAII, is not patient dependent balanced pH in the tumor microenvironment that may be in
but rather associated with the oncogenic process. This is in line turn favorable for the maintenance of immune effector
with our observation in melanoma that immune responsive- mechanisms. Indeed, CAIX is a hypoxia-inducible gene that
ness is predetermined before treatment and seems related participates in the VHL/hypoxia pathway (20, 21). Interestingly,
primarily to the biology of the tumors rather than the genetic unsupervised analysis of genes found up-regulated in RCC
background of the patient (10). Interestingly, in the same suggested that CAIX expression is tightly coordinated with that
study, none of seven metastases of melanoma, another cancer of N-myc and signal transducers and activators of transcription
highly responsive to systemic IL-2 administration, were found (STAT) interactor (NMI; Fig. 1). This gene interacts with N-myc,
to express CAIX. This finding was confirmed more recently in a C-myc, and other zipper transcription factors (22). Interest-
larger data set where 59 melanoma lesions were compared with ingly, NMI interacts with all STATs except STAT2 and
>100 tumors of other histology (16). This surprising finding augments STAT-1 – mediated transcription in response to IL-2,
suggests that the upstream pathways leading to immune IFN-a, IFN-h, and IFN-g signaling.1 In particularly, the ability
response in melanoma may be different from that of RCC if of NMI to interact with N-myc may be of particular significance
CAIX bears more than a coincidental association with immune because we have previously noted that N-myc was up-regulated
responsiveness. It is also possible that CAIX acts upstream of a in a melanoma metastasis responding to IL-2 therapy (7). In
possible common denominator pathway for immune rejection addition, CAIX expression was tightly associated with other
of melanoma and RCC tumors. We hypothesize that the final IFN-responsive genes including IFRG28, IFI16, and C-myc and
pathway leading to tumor rejection may be similar between the several transcription factors. This information suggests that
two cancers and in fact may bear similarities to the biological CAIX may either play a metabolic role broader than simply
pathway leading to rejection of basal cell carcinoma by toll maintaining extracellular ion homeostasis and may facilitate
receptor agonists (17) which seem synergistic with IL-2 (18). In some critical functions associated with acute inflammatory
addition, similarities seem to exist between IL-2 – induced processes or its expression may depend on an active immune
tumor regression (7) and the mechanism(s) inducing allograft environment as a downstream result of hypoxia-inducible
rejection (8) or maintenance of systemic lupus erythematosus factor-1a activation (23). In this regard, it is interesting to note
(19). All these biological phenomena seem to depend on the
convergence in the target organ of a final pathway that seems
characterized by the activation of innate immune effectors
1
through induction of type one IFNs (2, 7, 8, 17, 19). Stroncek et al., submitted for publication.
Clin Cancer Res 2005;11(10) May 15, 2005 3602 www.aacrjournals.org
CAIX and Immune Responsiveness
the parallel expression of CCAAT/enhancer binding protein (http://cellspace.cellomics.com/CellSpace/csmanager.asp).
(C/EBP) in the CAIX cluster, because this protein is dependent Therefore, extensive biological characterization of these genes
on STAT-3 signaling (24) induced by oxidative stress in parallel function will need to be entertained in the future to better
to the expression of hypoxia-inducible factor-1a (25) respon- understand its relationship with immune responsiveness.
sible for the expression of CAIX (23). Therefore, it is possible It is also possible that CAIX may be target of cellular or
that in VHL-defective tumors, even in normoxic conditions, a humoral immune responses because it is a transmembrane
pseudohypoxic state is maintained which leads to the protein sensitive to antibody-dependent cytotoxicity (26). This
expression of transcription factors with downstream constitu- may be a significant mechanism because CAIX expression is
tive expression of CAIX, C/EBP, NMI, c-myc, and other genes variably modulated by cytokines released in vivo in response to
identified in this signature. This is a likely possibility and we do systemic IL-2 administration (9) such as IL-2 itself and IFNs
agree with the authors’ statement that ‘‘CAIX may serve as a which enhance and IL-1h and IL-4 which decrease its
surrogate marker for some other critical hypoxia-inducible expression (27). However, thus far no evidence of natural
factor – mediated protein (perhaps NMI) that is more directly cellular or humoral responses against CAIX have been reported.
associated with immune responsiveness (2).’’ The reason why Finally, it could be argued that if tumor immune respon-
this phenomenon would lead to enhanced responsiveness siveness is pre-determined, non – immune-responsive tumors
remains, however, elusive. A preliminary quest for functional should not be treated. Obviously, this will depend on the
associations among genes in the CAIX cluster based on the mechanisms leading to immune responsiveness and whether
database for annotation visualization and integrated discovery they could be manipulated in vivo. This could be only
(DAVID at http://david.niaid.nih.gov/david/upload.asp) could determined when a better understating of the algorithm
not identify known direct interactions between CAIX and other determining immunologically mediated tumor rejection will
genes included in the signature. Similarly, CellSpace-based be achieved. CAIX represents another important step in that
mining did not identify strong evidence of known associations direction.
References
1. Atkins MB, Regan M, McDermott D, et al. Carbonic 10. Wang E, Miller LD, Ohnmacht GA, et al. Prospective 20. Ivanov SV, Kuzmin I, Wei MH, et al. Down-
anhydrase IX expression predicts outcome in interleu- molecular profiling of subcutaneous melanoma metas- regulation of transmembrane carbonic anhydrases
kin-2 therapy of renal cancer. Clin Cancer Res 2005;11: tases suggests classifiers of immune responsiveness. in renal cell carcinoma cell lines by wild-type
3714 ^ 21. Cancer Res 2002;62:3581 ^ 6. von Hippel-Lindau transgenes. Proc Natl Acad
2. Atkins MB, Regan M, McDermott D. Update on the 11. Panelli MC, Martin B, Nagorsen D, et al. A genomic Sci U S A 1998;95:12596 ^ 601.
role of interleukin 2 and other cytokines in the treat- and proteomic-based hypothesis on the eclectic 21. Ivanov S, Liao SY, Ivanova A, et al. Expression of
ment of patients with stage IV renal carcinoma. Clin effects of systemic interleukin-2 administration in the hypoxia-inducible cell-surface transmembrane car-
Cancer Res 2004;10:6342 ^ 6S. context of melanoma-specific immunization. Cells Tis- bonic anhydrases in human cancer. Am J Pathol 2001 ;
3. Phan GQ, Attia P, Steinberg SM,White DE, Rosenberg sues Organs 2003;177:124 ^ 31. 158:905 ^ 19.
SA. Factors associated with response to high-dose in- 12. Monsurro’ V, Wang E, Panelli MC, et al. Active- 22. Zhu M, John S, Berg M, Leonard WJ. Func-
terleukin-2 inpatients with metastatic melanoma. JClin specific immunization against cancer: is the problem tional association of Nmi with Stat5 and Stat1 in
Oncol 2001 ;19:3477 ^ 82. at the receiving end? Semin Cancer Biol 2003;13: IL-2- and IFNg-mediated signaling. Cell 1999;96:
4. Royal RE, Steinberg SM, Krouse SE, et al. Correlates of 473 ^ 80. 121 ^ 30.
response to IL-2 therapy in patients treated for meta- 13. Monsurro’ V, Wang E, Yamano Y, et al. Quiescent 23. Grabmaier K, Weijer t MCAd, Verhaegh GW,
static renal cancer and melanoma. Cancer J Sci Am phenotype of tumor-specific CD8+ T cells following Schalken JA, Oosterwijk E. Strict regulation of
1996;2:91 ^ 8. immunization. Blood 2004;104:1970 ^ 8. CAIX(G250/MN) by HIF-1ain clear cell renal cell car-
5. White RLJ, Schwartzentruber D, Guleria AS, et al. 14. Bui MH, Seligson D, Han KR, et al. Carbonic anhy- cinoma. Oncogene 2004;23:5624 ^ 31.
Cardiopulmonary toxicity of treatment with high-dose drase IX is an independent predictor of survival in ad- 24. Numata A, Shimoda K, Kamezaki K, et al. Signal
interleukin-2 in 199 consecutive patients with meta- vanced renal clear cell carcinoma: implications for transducers and activators of transcription 3 augments
static melanoma or renal cell carcinoma. Cancer 1994; prognosis and therapy. Clin Cancer Res 2003;9: the transcriptional activity of CCAAT/enhancer bind-
74:3212 ^ 22. 802 ^ 11. ing protein a in G-CSF signaling pathway. J Biol Chem
6. Webster AC, Playford EG, Higgins G, Chapman JR, 15. Wang E, Lichtenfels R, Bukur J, et al. Ontogeny and 2005.
Craig J. Interleukin 2 receptor antagonists for kidney oncogenesis balance the transcriptional profile of renal 25. Tacchini L, Fusar-Poli D, Bernelli-Zazzera A. Activa-
transplant recipients. Cochrane Database Syst Rev cell cancer. Cancer Res 2004;64:7279 ^ 87. tion of transcription factors by drugs inducing oxida-
CD003897 2004. 16. Wang E, Panelli MC, Zavaglia K, et al. Melanoma- tive stress in rat liver. Biochem Pharmacol 2002;63:
7. Panelli MC,Wang E, Phan G, et al. Genetic profiling of restricted genes. J Transl Med 2004;2:34. 139 ^ 48.
peripheral mononuclear cells and melanoma metasta- 17. Sauder DN. Imiquimod: modes of action. Br J Der- 26. Liu Z, Smyth FE, Renner C, Lee FT,
ses in response to systemic interleukin-2 administra- matol 2003;149:5 ^ 8. Oosterwijk E, Scott AM. Anti-renal cell carci-
tion. Genome Biol 2002;3:RESEARCH0035. 18. Bong AB, Bonnekoh B, Franke I, Schon MP, Ulrich noma chimeric antibody G250: cytokine en-
8. Sarwal M, Chua MS, Kambham N, et al. Molecular J, Gollnick H. Imiquimod, a topical immune response hancement of in vitro antibody-dependent
heterogeneity in acute renal allograft rejection identi- modifier, in the treatment of cutaneous metastases cellular cytotoxicity. Cancer Immunol Immun-
fied by DNA microarray profiling. N Engl J Med 2003; of malignant melanoma. Dermatology 2002;205: other 2002;51:171 ^ 7.
349:125 ^ 38. 135 ^ 8. 27. BlokVT, Gelderman KA,Tijsma OH, Daha MR, Gorter
9. Panelli MC, White RL Jr, Foster M, et al. Forecasting 19. Banchereau J, PascualV, Palucka AK. Autoimmunity A. Cytokines affect resistance of human renal tumour
the cytokine storm following systemic interleukin-2 through cytokine-induced dendritic cell activation. Im- cells to complement-mediated injury. Scand J Immu-
administration. J Transl Med 2004;2:17. munity 2004;20:539 ^ 50. nol 2003;57:591 ^ 9.
www.aacrjournals.org 3603 Clin Cancer Res 2005;11(10) May 15, 2005
Related docs
