Download Multidrug Resistance in Breast Cancer—Is the Jury in Yet?

EDITORIALS
Multidrug Resistance in Breast Cancer—
Is the Jury in Yet?
S. B. Kaye*
The term ‘‘multidrug resistance’’ has come to mean different
things to different people. Initially, it was coined to describe the
laboratory phenomenon of cross-resistance to various natural
products, and this phenomenon was later found to relate to increased expression of P-glycoprotein (gp170) (1). Since then,
many laboratory scientists and some clinicians have mistakenly
assumed that the clinical observation of drug resistance to a wide
range of cytotoxic agents has the same basis. Are the two in fact
related? In this issue of the Journal, Trock et al. (2) have conducted an exhaustive review of the literature in an attempt to
answer this question as it relates to breast cancer. There has been
a plethora of studies with widely varying conclusions, and an
attempt to distill them into a single review article is indeed
praiseworthy. But there are problems.
First, and predictably, the use of meta-analysis in this field is
bound to provoke criticism, and the authors acknowledge this.
While clinicians have readily accepted for some years the value
of meta-analysis in allowing conclusions to be reached when
randomized trial data are pooled (3), using it to look for clinical
associations of laboratory observations is fraught with danger.
Trock et al. have meticulously drawn up criteria for judging the
quality of the various studies; they proceed to point out that very
few studies actually meet these criteria. An important example is
the fact that 67% of the studies that were based on immunohistochemical detection of gp170, which is probably the preferred
technique, used only one antibody, despite the acknowledged
major problem of cross-reactivity with other proteins. Does this
high level of concern about the validity of the basic data not call
into question the reliability of any conclusions drawn from
them?
Second, no attempt was made to analyze in detail the various
criteria for ‘‘positivity,’’ although Trock et al. did suggest that
such differences probably contribute significantly to heterogeneity. Is there, for example, some threshold percentage of positively staining cells with the use of immunohistochemistry?
Since, as stated by Trock et al. (2), ‘‘approximately 40% of all
breast tumors express detectable levels of MDR1/gp170,’’ it
would have been helpful to have had some notion of the range
of positivity involved.
Third, and most importantly, there are the tentative conclusions by Trock et al. regarding the association between clinical
response and gp170 positivity. It is clear from the article by
Trock et al. that exposure to cytotoxic drugs (not exclusively of
the MDR1-related type) increases gp170 expression [two-sided
902
EDITORIALS
P<.0001, Table 4 in (2)]. This is entirely in keeping with laboratory observations on induction of multidrug resistance (4).
Does this represent induced expression, or does it result from the
selection of tumor subpopulations based on failure of response
to treatment? Trock et al. favor the latter explanation, stating that
gp170 is likely to have a role in conferring clinical resistance ‘‘in
a significant proportion of breast tumors.’’
One problem in assessing this proposal is the major drawback
that very few data are available on serial samples from the same
patient, and Trock et al. rightly urge that such studies be done in
the future. Nevertheless, there remains the intriguing observation of a highly significant association between gp170 positivity,
particularly in samples taken after chemotherapy, and the lack
of response to that treatment. This association is seen after chemotherapy with many drugs as well as with those restricted to
the MDR1 family, with a two-sided P value on the association of
<.0001 [Table 6 in (2)]. On the other hand, and importantly,
Trock et al. observed that gp170 positivity before chemotherapy
is not significantly associated with lack of response to subsequent treatment, with a P value of .088 on the relative risk, albeit
from a pooled total of only 115 case patients.
Thus, the conclusion made by Trock et al. is open to debate.
Despite all the methodologic reservations expressed above, there
does appear to be an association between poor clinical response
and gp170 positivity in samples from treated breast cancer patients. However, this finding does not necessarily mean that the
two are causally linked. The development of drug resistance is
likely to lead to tumor cell populations with various molecular
and genetic characteristics, for example, increased numbers of
cells with mutations in the p53 gene. MDR1 gene expression is
regulated by several factors; experimentally, these factors have
been shown to include mutant p53, which specifically stimulates
the MDR1 promoter (5). Thus, an alternative explanation for
several of the observations made by Trock et al. is that gp170
positivity is an epiphenomenon and that, in poorly responsive
cases of breast cancer, it is a result of other genetic changes.
Such an argument is strengthened by the absence of a significant
association between initial gp170 positivity and response to
treatment; if there were a strong functional relationship, one
*Correspondence to: S. B. Kaye, M.D., CRC Department of Medical Oncology, University of Glasgow, CRC Beatson Laboratories, Garscube Estate,
Switchback Rd., Bearsden, Glasgow G61 1BD, U.K.
© Oxford University Press
Journal of the National Cancer Institute, Vol. 89, No. 13, July 2, 1997
would expect there to have been a positive conclusion, as has
been observed, for example, in studies linking the clinical resistance to platinum compounds to the presence of mutant p53
sequences in primary tumor samples of ovarian cancer (6,7).
And yet much still has to be learned. There remains the
tantalizing observation made by some authors of a significant
association between gp170 positivity and survival in a range of
tumor types. Although this association is not confirmed in the
article by Trock et al., a testable hypothesis is that gp170 expression is a marker of cell behavior and has a possible role in
the processes of invasion and metastasis (8). Of course, clinicians are impatient people, and they have already embarked on
a number of trials of multidrug resistance modulators, based on
positive experimental data. The findings from these trials have
largely been disappointing, at least with regard to solid tumors,
but some would say that, in many of these trials, the use of
modulators with little potency is an important consideration.
Recently, enthusiasts for multidrug resistance modulation have
been stimulated by the intriguing experimental observation that
exposure of tumor cells in vitro to the modulating agent PSC 833
suppresses the emergence of MDR1 gene mutants, thereby inhibiting the development of drug-resistant cells, in the presence
of the MDR1-related drug doxorubicin (9). This observation
raises potentially important clinical issues. In particular, if one
assumes that gp170 is indeed one of the factors responsible for
clinical drug resistance to MDR1-related drugs, would the time
to interfere with its function not best be during the initial exposure rather than later in the patient’s clinical history, as Trock et
al. propose, by which stage it is even more likely that other
factors will dominate the picture?
Trock et al. conclude their worthy review article by express-
Journal of the National Cancer Institute, Vol. 89, No. 13, July 2, 1997
ing their hope that further discussion will take place. This should
indeed be the case, with maximum dialogue between laboratory
and clinical scientists. Until then, the jury on gp170 in breast
cancer and in other cancers remains firmly out.
References
(1) Kartner N, Riordan JR, Ling V. Cell surface P-glycoprotein associated with
multidrug resistance in mammalian cell lines. Science 1983;221:1285-8.
(2) Trock BJ, Leonessa F, Clarke R. Multidrug resistance in breast cancer: a
meta-analysis of MDR1/gp170 expression and its possible functional significance. J Natl Cancer Inst 1997;89:917-31.
(3) Systemic treatment of early breast cancer by hormonal, cytotoxic or immune
therapy. 133 randomised trials involving 31,000 recurrences and 24,000
deaths among 75,000 women. Early Breast Cancer Trialists’ Collaborative
Group. Lancet 1992;339:1-15;71-85.
(4) Chaudhary PM, Roninson IB. Induction of multidrug resistance in human
cells by transient exposure to different chemotherapeutic drugs. J Natl Cancer Inst 1993;85:637-9.
(5) Chin KV, Ueda K, Pastan I, Gottesman MM. Modulation of activity of the
promoter of the human MDR1 gene by Ras and p53. Science 1992;255:45962.
(6) Righetti SC, Della Torre G, Pilotti S, Menard S, Ottone F, Colnaghi MI, et
al. A comparative study of p53 mutations, protein accumulation, and response to cisplatin-based chemotherapy in advanced ovarian carcinoma.
Cancer Res 1996;56:689-93.
(7) Buttitta F, Marchetti A, Gadducci A, Pellegrini S, Morganti M, Carnicelli V,
et al. p53 alterations are predictive of chemoresistance and aggressiveness in
ovarian carcinomas: a molecular and immunohistochemical study. Br J Cancer 1997;75:230-5.
(8) Pinedo HM, Giaccone G. P-glycoprotein—a marker of cell behavior [editorial]. N Engl J Med 1995;333:1417-9.
(9) Beketic-Oreskovic L, Duran GE, Chen G, Dumontet C, Sikic BI. Decreased
mutation rate for cellular resistance to doxorubicin and suppression of mdr1
gene activation by the cyclosporin PSC 833. J Natl Cancer Inst 1995;87:
1593-602.
EDITORIALS
903