The report of Godal et al1  raises an important issue on the effectiveness of apoptosis inhibitors in preventing death induced by cytotoxic lymphocytes (CLs) in clinically relevant settings. A large variety of tumors express apoptosis inhibitors such as Bcl-2 and PI-9, which protect cells from granzyme B (GrB)–induced apoptosis in in vitro settings,2-4  and were thus proposed to induce immune escape of tumors. Based on in vitro assays Godal et al1  now claim that expression of these apoptotic regulators is irrelevant for CL sensitivity of lymphomas and therefore unlikely to affect immunotherapy of lymphomas. However, a major limitation of this conclusion is that it is based on in vitro analysis, and current data suggest that this doesn't necessarily represent the in vivo situation.

CL degranulation induces apoptosis mainly via GrB, while other granzymes bring about different modes of cell death. As PI-9 inhibits GrB, it effectively prevents apoptosis, but not these other deaths. In agreement, Godal et al show that in vitro cytotoxicity of lymphomas does not depend on PI-9 or Bcl-2 expression, which then warrants the conclusion that these do not protect lymphomas. Although relevant, this conclusion is not supported by data from other groups. For instance, Classen et al5  concluded that PI-9 expression in pediatric acute lymphoblastic leukemias is correlated with protection in vitro. Similarly, PI-9 expression in MCF-7 cells prevents death induced by long-term activated natural killer (NK) cells (expressing little GrB), while it is ineffective against shortly activated NK cells (expressing high levels of GrB).2  These findings suggest that PI-9 efficacy depend on the conditions used. Although this may appear trivial, Godal et al1  assume that cytotoxicity measured in their in vitro assays is identical to the in vivo situation. This strongly contrasts with current knowledge, as it is clear that CLs isolated from blood contain less granzymes than most CL lines. For instance, granzyme M, which induces an alternative death pathway, is not detected in activated cytotoxic T lymphocytes (CTLs), while it is present in CTL lines.6,7  Similarly, NK cells have little GrB, while lymphokine-activated killer (LAK) cells contain enormous amounts.6  It is therefore likely that the level of CL activation, and thus the diversity of granzymes expressed, determines the efficacy of inhibition by PI-9. In agreement, we have shown that expression of SPI-6 (mouse PI-9) in T lymphomas does not protect these cells against cytotoxicity induced by a CL in vitro, but does convey a level of protection in vivo.8  Also, the clinical data available support this notion. The group of Oudejans and Kummer provided compelling evidence that PI-9 expression is associated with poor prognosis in anaplastic large B-cell lymphoma.9  More importantly, they showed that PI-9 expression is an important determinant in disease-free survival time of melanoma patients following immunotherapy.10 

In conclusion, measuring cytotoxicity using highly activated CL in vitro may over-exaggerate the effectiveness of these cells and thus underestimate the protective capacity of antiapoptotic molecules. Indeed, the experimental evidence pictures PI-9 as a crucial determinant in the outcome of immunotherapeutic approaches in cancer.

Is there any evidence supporting a critical role of ectopic PI-9 expression in tumor immune escape?

The authors appreciate the interest of Bots et al in our recent paper regarding the role of PI-9 in CTL- and NK-cell–mediated lysis of lymphoma cells.1  The hypothesis that the granzyme B inhibitor PI-9 expressed in certain human malignancies may prevent their clearance by the immune system has been controversial for many years.2  We think, however, that the argumentation of Bots et al is debatable. The study by Classen et al,3  which Bots et al cite as not supporting our results, has investigated leukemic blasts from 2 patients only. The notion that “cytotoxic lymphocytes isolated from peripheral blood contain less granzymes than most cytotoxic lymphocyte lines,” referring to the study by Sayers et al,4  is partially true for granzyme M, but not for granzyme B, which is the only granzyme inhibited by PI-9. Sayers et al clearly show that the 2 major granzymes A and B are highly expressed in freshly isolated CD8+ T cells at a level similar to cytotoxic lymphocyte lines. In their own study, Medema et al5  have actually shown that indeed, transfection of SPI-6 (mouse PI-9) into tumor cells conveyed a level of protection from granzyme B–induced apoptosis, but did not inhibit CTL-induced cytolysis. This finding is in accordance with various more recent studies showing that other granzymes with strong antitumor activity can substitute granzyme B in vitro and in vivo, as reviewed by Lieberman.6  In the 2 studies mentioned by Bots et al, which linked PI-9 expression in tumor tissues to clinical outcome, there was no evidence of a clinically effective immune response to tumor in which granzyme B resistance might play a role.7,8 

In our opinion, the most convincing in vivo studies regarding the potential role of granzyme B and perforin in cancer are those in knock-out mice reported by Trapani's group.9,10  Lymphoma cells grew efficiently in perforin-deficient mice, whereas granzyme A– and/or B–deficient mice rejected large tumor doses as avidly as wild-type mice, indicating that granzyme B was completely dispensable in lymphoma eradication.9  These studies also clearly show a strong correlation between cytolytic activity of granzyme B knockout CTLs in vitro and in vivo.10  Taken together, the studies by the groups of Trapani and Lieberman6,9,10  are in line with our studies in human lymphoma cells as they provide strong evidence that granzyme B is not critical for antitumor effector functions in vitro and in vivo. We therefore believe that the results from the murine knock-out models as well as our human in vitro experiments justify the conclusion considering lymphoma to be sensitive to perforin-dependent pathways despite PI-9 expression. Of course, we fully agree with Bots et al that the final answer regarding the role of PI-9 in immune rejection of human tumors can only be provided in vivo, which would require clinical studies based on immunotherapy with proven clinical efficacy.

Correspondence: Carmen Scheibenbogen, Hematology and Oncology, Campus Benjamin Franklin, Charité, Hindenburgdamm 30, D-12200 Berlin, Germany; e-mail: carmen.scheibenbogen@charite.de.

1
Godal R, Keilholz U, Uharek L, et al. Lymphomas are sensitive to perforin dependent cytotoxic pathways despite expression of PI-9 and overexpression of bcl-2.
Blood
.
2006
;
107
:
3205
-3211.
2
Trapani JA, Sutton VR. Granzyme B: pro-apoptotic, antiviral and antitumor functions.
Curr Opin Immunol
.
2003
;
15
:
533
-543.
3
Classen CF, Ushmorov A, Bird P, Debatin KM. The granzyme B inhibitor PI-9 is differentially expressed in all main subtypes of pediatric acute lymphoblastic leukemias.
Haematologica
.
2004
:
89
:
1314
-1321.
4
Sayers TJ, Brooks AD, Ward JM, et al. The restricted expression of granzyme M in human lymphocytes.
J Immunol
.
2001
;
166
:
765
-771.
5
Medema JP, de Jong J, Peltenburg LT, et al. Blockade of the granzyme B/perforin pathway through overexpression of the serine protease inhibitor PI-9/SPI-6 constitutes a mechanism for immune escape by tumors.
Proc Natl Acad Sci U S A
.
2001
;
98
:
11515
-11520.
6
Lieberman J. The ABCs of granule-mediated cytotoxicity: new weapons in the arsenal.
Nat Rev Immunol
.
2003
;
3
:
361
-370.
7
ten Berge RL, Meijer CJ, Dukers DF, et al. Expression levels of apoptosis-related proteins predict clinical outcome in anaplastic large cell lymphoma.
Blood
.
2002
;
99
:
4540
-4546.
8
Van Houdt IS, Oudejans JJ, van den Eertwegh AJ, et al. Expression of the apoptosis inhibitor protease inhibitor 9 predicts clinical outcome in vaccinated patients with stage III and IV melanoma.
Clin Cancer Res.
2005
;
11
:
6400
-6407.
9
Smyth MJ, Street SE, Trapani JA. Cutting edge: granzymes A and B are not essential for perforin-mediated tumor rejection.
J Immunol
.
2003
;
171
:
515
-518.
10
Davis JE, Smyth MJ, Trapani JA. Granzyme A and B-deficient killer lymphocytes are defective in eliciting DNA fragmentation but retain potent in vivo antitumor capacity.
Eur J Immunol
.
2001
;
31
:
39
-47.
1
Godal R, Keilholz U, Uharek L, et al. Lymphomas are sensitive to perforin dependent cytotoxic pathways despite expression of PI-9 and overexpression of bcl-2.
Blood
.
2006
;
107
:
3205
-3211.
2
Bird CH, Sutton VR, Sun J, et al. Selective regulation of apoptosis: the cytotoxic lymphocyte serpin proteinase inhibitor 9 protects against granzyme B-mediated apoptosis without perturbing the Fas cell death pathway.
Mol Cell Biol
.
1998
;
18
:
6387
-6398.
3
Sutton VR, Vaux DL, Trapani JA. Bcl-2 prevents apoptosis induced by perforin and granzyme B, but not that mediated by whole cytotoxic lymphocytes.
J Immunol
.
1997
;
158
:
5783
-5790.
4
Bots M, van Bostelen L, Rademaker TG, et al. Serpins prevent granzyme-induced death in a species-specific manner.
Immunol Cell Biol
.
2006
;
84
:
79
-86.
5
Classen CF, Ushmorov A, Bird P, Debatin KM. The granzyme B inhibitor PI-9 is differentially expressed in all main subtypes of pediatric acute lymphoblastic leukemias.
Haematologica
.
2004
:
89
:
1314
-1321.
6
Sayers TJ, Brooks AD, Ward JM, et al. The restricted expression of granzyme M in human lymphocytes.
J Immunol
.
2001
;
166
:
765
-771.
7
Bots M, Kolfschoten IG, Bres SA, et al. SPI-CI and SPI-6 cooperate in the protection from effector cell-mediated cytotoxicity.
Blood
.
2005
;
105
:
1153
-1161.
8
Medema JP, de Jong J, Peltenburg LT, et al. Blockade of the granzyme B/perforin pathway through overexpression of the serine protease inhibitor PI-9/SPI-6 constitutes a mechanism for immune escape by tumors.
Proc Natl Acad Sci U S A
.
2001
;
98
:
11515
-11520.
9
ten Berge RL, Meijer CJ, Dukers DF, et al. Expression levels of apoptosis-related proteins predict clinical outcome in anaplastic large cell lymphoma.
Blood
.
2002
;
99
:
4540
-4546.
10
Van Houdt IS, Oudejans JJ, van den Eertwegh AJ, et al. Expression of the apoptosis inhibitor protease inhibitor 9 predicts clinical outcome in vaccinated patients with stage III and IV melanoma.
Clin Cancer Res
.
2005
;
11
:
6400
-6407.