Individual & Population Tailored Treatment Strategies For Cancer

Time and again medical specialists and cancer researchers have been confronted with the question – why is a certain patient with a known curable cancer not responding to conventional treatment or why did he suffer a relapse?

The simplest explanation for such differences must be that either the “unresponsive” cancer is in some way biologically different from the responsive cancer, or that the treatment given is not as effective in the patient who does badly, because of differences, for example, in the amount of effective drug that reaches the tumor. Perhaps there is no single answer to these questions and multiple factors contribute. Whatever the complexities, the variations in the biology, natural course and treatment of cancer are intriguing and challenging. Dissection of the causes of such variations will eventually provide the tools to offer curative regimens to individual patients, rather than using treatment derived from group studies, which give the best overall results, but which are clearly not optimal for all patients in the group. What we call cancer is in reality a group of diseases, each with one common property - the abnormal, and unbridled proliferation of cells which have escaped from normal control. With time, these abnormal cells also gain the potential to spread and survive in diverse tissue environments, i.e. to “metastasize” to various parts of the body.

Kishor Bhatia, Director of the Research Center, King Fahd National Children’s Cancer Center, and of INCTR’s Translational Research Program.

Over the years, cancers have mostly been diagnosed and treated on the basis of their clinical and morphological features. Conventional treatments of cancer have revolved around the modalities of surgery, radiation and chemotherapy, used singly or in combination, according to clinical protocols. Treatment strategies for cancer have, until recently, been guided by experience based on morphological diagnosis and clinical trials of therapy conducted mostly in North America, Europe and Australia.

It became apparent that clinical (e.g. disease extent) and morphologic (i.e the appearance under the microscope) features could be used to classify cancers and to predict the likelihood of success with a given therapy. As we have gained additional tools to classify cancer, using for instance antibodies to identify variations in the expression of a limited panel of proteins, our ability to define “less responsive” malignancies among otherwise morphologically and clin.ically identical cancers has improved. Such analyses provided the preliminary insights that expression (or lack of expression) of specific genes in morphologically identical cancers may allow prediction of treatment failure. Since the behavior of a given cancer is dependent directly upon the profile of expression of its genes, it is likely that defining the expression profile of cancers will further sharpen classification, making it even more clinically relevant. Furthermore, it is now clear that cancers are ultimately the results of molecular genetic changes in normal cells that are responsible for altering the pattern of expression of proteins, and therefore creating the type of cellular behavior that we refer to as malignant (e.g. spreading to other parts of the body, constant accumulation of cells).

Rong Bu, a Chinese physician/scientist whose training in the King Fahd laboratory was initially supported by INCTR.

Such strategies should allow the oncologist to choose regimens that are most appropriate for the genetic profile of the cancer, irrespective of the morphology. Should we assume that molecularly identical cancers in different patients within defined geographic populations respond similarly to a chosen treatment plan? There is increasing evidence that in addition to variation in the genetic profile of the tumor, the patient’s own genetic make-up will also influence clinical outcome, be it response toxicity or long-term sequelae. Can we then presuppose that the treatment approaches standardized in one geographic population are ideal for other populations? Unfortunately there is evidence that would suggest that biological, and ultimately, molecular, differences may exist in morphologically identical cancers in diverse populations. This should provide the impetus to conduct studies on whether the fraction of clinically relevant subclasses differ geographically, since the overall response to therapy in any population will be influenced by the composition of the prevalent subclasses that are predictive of treatment outcome.

In the future, therapeutic strategies are more likely to be guided by study of the “molecular lesions“ that cause cancer. Novel drugs are being developed that are specifically targeted against such lesions. Thus for devising can.cer management plans, it may no longer be sufficient to define, for example, the total burden of breast cancer, but to assess the expression in the tumor cells of a particular protein (e.g “neu”, a gene amplified in a subset of breast cancer, which, therefore, can be treated by strategies specifically targeting the expression of “neu”). It is clear that a shift is occurring from the use of systemic cytotoxic agents to tumor- specific therapies, and the practice of directly adopting treatments standardized in the West may need to be revised.

Another reason to revisit the above issue stems from gathering evidence that inter-individual variability in the metabolism of drugs can significantly modify their bio-availability, efficacy and toxicity. Even within defined geographic patient populations and using more or less uniform approaches to treatment and support of cancer, variation in clinical outcome has been documented.

In a recently published report from the US, racial and ethnic differences were noted in the survival of children with acute lymphoblastic leukemia treated with contemporary risk-based therapy ^[1]). Multivariate analysis showed that Asian children in the US had a better outcome in terms of survival compared with white, followed by Hispanic and black children. The authors speculate that one of the factors that would have contributed to the differential outcome is genetic variation in the metabolism of various chemotherapeutic drugs that exist in the study population, based on ethnic and racial lines. Several studies have shown that polymorphisms for genes responsible for drug metabolism influence drug disposition in the body. Genetic polymorphisms are defined as a variation (i.e a specific change in the coding sequence for the gene, rather like the misspelling of a word) in the DNA sequence at a given nucleotide position that is present in a small percentage (e.g., 1%) of the population (Fig. 1). These are called single nucleotide polymorphisms, or SNPs. There are more than three million SNPs described in the human genome.

LEFT PANEL Agarose gel electrophoresis of PCR products from a multiplex reaction. The presence of GSTT1 is shown in lanes 6 and 7 (480bp product). GSTM1 gene is seen in lanes 1 and 2 (230bp product).Beta-globin gene is co-amplified as an internal control (110bp band). Lanes 3, 4 and 5 represent null genotypes. B = blank. RIGHT PANEL Agarose gel electrophoresis of Hinfl- digested PCR products. MTHFR677C allele yields a 198bp band and MTHFR677T allele yields 175bp band (+23bp band not detected). Lanes a and b represent CC wild type genotype, lanes c and d represent CT heterozygotes and lanes e and f,TT mutant homozygotes. M = molecular weight marker (sizes in bp are shown). Figure 1. SNP analyses

There are also recent reports that drugs used in supportive care of cancer patients are also handled differently by individuals ^[4]). Those who have deletions or mutations of cytochrome P-450 enzymes have different responses to anti-emetic drugs like ondansetron. Thus the ultra-rapid metabolizers of the drug have more significant nausea and vomiting. SNPs have lately also been described for genes that encode erythropoetin and other colony-stimulating factors.

In the future, pharmacogenomics or pharmacogenetics will dictate which drug is best, and in what dose in a given individual, based on his/her genotypic profile. Integration of pharmacogenetic data into treatment strategies for cancer and supportive care is only a matter of time. Successful treatment of cancer will not only depend upon histological identification but also on complete molecular subtyping of the tumor and genotyping of the patient. It is also imperative that we move from mere adaptation of protocols/therapies established in the Euro-American arena to conducting clinical trials within defined geographic populations and selecting optimal treatment strategies for the combination of the molecular subtype of a cancer and the genotype of the population being treated.

Kishor Bhatia¹ & Rajeev Sathiapalan²

References:

^[1] Smita Bhatia, Harland N. Sather, Nyla A. Heerema, Michael E. Trigg, Paul S. Gaynon, and Leslie L. Robison. Racial & Ethnic differences in the survival of children with acute lymphoblastic leukemia. Blood 100: 1957-1964, September 2002.