Breast-feeding, HAMLET, and cancer

There is epidemiological evidence that breast-feeding protects against tumor development in children. An initial report from Davis et al showed that the overall incidence of tumors in children up to the age of 15 was lower in children who had been breast-fed than among those who had been formula-fed (Davis et al., 1988). This effect was especially pronounced for lymphomas. A number of case-control studies have since shown a clear protection against development of Hodgkin's disease but with variable results for non-Hodgkin's lymphoma, acute lymphoblastic leukemia or other cancer forms (Davis et al, 1998 and references therein). Only one study has investigated the duration of breast-feeding in relation to tumor disease (Mathur et al., 1993); it showed a protection against all cancers with the strongest protection against lymphomas. The tumor cases were breast-fed 8 months compared with 10 months for controls. The difference was even more pronounced when exclusive breast-feeding was investigated. Here tumor cases were exclusively breast-fed for an average of 3.2 months compared with 4.6 months for controls.
Breast-feeding has also been reported to protect against development of immunological diseases, such as allergy (Ahmed & Fuchs, 1997; Saarinen & Kajosaari, 1995), infantile diabetes mellitus (Gerstein, 1994; Mayer et al., 1988) and inflammatory gastrointestinal disease (Koletzko et al., 1989). These results imply that factors in milk may help optimise the function of the immune system and regulate cell populations that results in les risk of tumor development and less risk for abberant immunity in the rapidly growing neonatal intestinal tract.

Besides HAMLET (see below) studies have shown that lactoferrin can inhibit solid tumor growth in mice (Bezault, 1994) and that unsaturated fatty acids that are present in human milk may be used to treat certain types of tumors in animal models or tumor cells in vitro (Zhu 1989, van Aswegen 1994).

In 1993, I joined the laboratory of Dr. Catharina Svanborg, Lund University, Sweden as a graduate student. During our studies of human milk and its effect on bacterial binding to epithelial cells we observed that a casein fraction of human milk not only blocked the binding of bacteria to the cells but also affected the viability of the lung cancer cell line used in the experiment (See Figure 1). We soon discovered, in collaboration with Dr Sten Orrenius and Dr Boris Zhivotovsky at Karolinska Institute in Stockholm, that the tumor cells were killed by apoptosis or programmed cell death, the type of cell death involved in elimination of faulty and superfluous cells from our tissues (Hakansson PNAS 1995). The effects were confined to human milk casein and were not found in any fraction from bovine, goat, or sheep milk.


Figure 1. Effect of a human milk fraction on the adherence of Streptococcus pneumoniae. (Top) Buccal epithelial cells are unaffected by the human milk fraction. (Bottom) A549 lung carcinoma cells become pyknotic and die after exposure to the milk fraction.

Identification and characterization of HAMLET

A post-doctoral fellow in the laboratory, Hemant Sabharwal, showed that the active fraction contained mainly alpha-lactalbumin (Figure 2) and we initially identified multimeric forms of alpha-lactalbumin that we thought were associated with the activity as the regular form (native, monomeric) of alpha-lactalumin had no tumoricidal activity.

A graduate student joining the laboratory, Malin Svensson, together with Sara Linse at the Department of Biophysical Chemistry, Lund University, showed that the activity did not require multimerization but was due to a conformational change in alpha-lactalbumin that it obtained after binding specific fatty acids found in human milk (Svensson JBC 1999). She was even able to make the anti-tumor form of the protein by first partially denaturing alpha-lactalbumin (by EDTA treatment) and then subjecting it to ion-exchange chromatography in the presence of the unsaturated fatty acids C18:1 and C18:2 (Svensson PNAS 2000). Based on this information Sara Linse came up with the acronym HAMLET for Human Alpha-lactalbumin Made LEthal to Tumor cells. I would also like to acknowledge the major contribution to these efforts made by our excellent laboratory technician Anki Mossberg.

Figure 3. Model of human alpha-lactalbumin with its four alpha-helices and two beta-sheets.

Further studies by Malin Svensson have found that the reason this activity is only found in human milk is not due to the alpha-lactalbumin. Bovine alpha-lactalbumin and alpha-lactalbumins from other species can become tumoricidal if associated with the correct fatty acids (Svensson Prot Sci 2003, Petterson BBRC 2006). Assocation with a fatty acid appears to be crucial as unfolded alpha-lactalbumin alone lacks tumoricidal activity (Svensson Prot Sci 2003). The specificity instead lays in the fatty acid. Only unsaturated C18 fatty acids in the cis conformation seem to provide the correct tumoricidal conformation of the protein (Svensson Prot Sci 2003b). C18 fatty acid binds in a compact conformation to the protein (Fast FEBS Lett 2005). For the latest information about current projects and publications about HAMLET structure and folding see the Svanborg web-site.

HAMLETs activity against tumor cells

HAMLET specifically kills tumor cells, while sparing healthy cells. To date we have tested up to 40 tumor cell lines and primary tumor cultures from different tissues as well as approximately 10 primary cell cultures from healthy tissues for their sensitivity to HAMLET. All tumor cells (including gliomas, adenocarcinomas from lung, breast, gastro-intestinal tract, urinary tract, and prostrate, fibrosarcomas and lymphoid and myeloid leukemias) have been found to be sensitive, whereas all healthy cells show resistance to the apoptosis-inducing activity of HAMLET.

HAMLET binds to both healthy (unsensitive) and tumor (sensitive) cells but can only enter tumor cells (Figure 3A). The reason for this tumor-specificity still eludes us. Together with Camilla Kohler (a graduate student in Sten Orrenius' lab in Stockholm), we found that HAMLET, once inside the cell, co-localized with mitochondria and that interaction with mitochondria resulted in depolarization of the mitochondrial membrane potential (Figure 3B), leading to what is called permeability transition, and release of mitochondrial factors that activates the apoptosis-specific proteases, the caspases (Hakansson ECR 1999, Kohler ECR 1999, and Kohler EJB 2000). HAMLETs activity on mitochondria was not confined to mitochondria from tumor cells, but could also be seen in mitochondria purified from healthy cells, suggesting that the main reason healthy cells are not sensitivie is that HAMLET is unable to enter the cells.

Mitochondrial activation results in the typical biochemical features of apoptosis including nuclear condensation and fragmentation (Figure 3C) and breakdown of DNA in specific fragment sizes (Figure 3D). HAMLET also localizes directly to the nuclei later in the process and can induce apoptotic DNA fragmentation directly in purified cell nuclei Hakansson ECR 1999).


Figure 3. (A) Binding of HAMLET (green) to the surface of A549 lung carcinoma cells (top) and subsequenct internalization of the protein (bottom). (B) HAMLET induces a depolarization of the mitochondrial membrane potential measured by a decreased intensity of the membrane potential dye TMRE (bottom). (C) HAMLET induces condensation and fragmentaiton of the nucleus of tumor cells. DNA is stained in the image with propidium iodide. (D). HAMLET induces typical high molecular fragmetnation of DNA indicative of apoptosis.

Caroline Duringer, another graduate student in the lab, showed that HAMLET binds strongly to the DNA associated histone H3 in vitro and that this interaction perturbed chromatin structure in vivo, which may contribute to HAMLETs apoptosis induction in tumor cells (Duringer, JBC 2003). This is supported by the fact that histone deacetylase inhibitors that also bind histones and destablize chromatin were able to synergize with HAMLET in producing DNA damage and cell death (Brest, Cancer Res 2007).

An interesting aspect of HAMLETs apoptosis inducing activity is that even though caspases are activated, they are not required for HAMLET to kill the cells (Hallgren, Apoptosis 2006). Similarly, the anti-apoptotic Bcl-2 proteins can not rescue the cells from HAMLET induced death and p53 does not appear to be involved in the nuclear signaling induced by chromatin pertubation. Thus HAMLET appears to be able to induce a caspase-independent cell death in tumor cells that is not associated with genes that are often mutated during tumor development. This could be a major advantage in the development of a treatment based on this protein.

HAMLETs effect against tumors in vivo

To date several studies have addressed the ability of HAMLET to kill tumor cells in vivo. Early attempts in mouse models were thwarted as we realized that serum (especially serum albumin) induces a partial inactivation of HAMLET based on binding to the fatty acid in HAMLET. The attempts that have since been done have addressed topical or mucosal administration of HAMLET.

Walter Fischer, then in Bergen, Norway, used HAMLET in a model where tumor tissue from patients with glioblastomas were transplanted under the skull of rats. In this model one dose of HAMLET was infused in the brain through a pump and tissue was tested for apoptosis induction and the rats were monitored for survival. HAMLET treatment delayed death of the rats and the most important finding was that when investigating the brain tissue they found that apoptosis was only induced in tumor cells, no healthy brain cells were affected (Fischer Cancer Res 2004). This indicated what we already know in vitro, that the protein only attacks tumor cells, and indicated that little side effects may be expected should this be used for treatment purposes. Walter, now in Copenhagen, Denmark, is currently planning a patient study in patients with glioblastomas.

In a second study, Lotta Gustafsson together with Irene Leijonhufvud conducted a study on benign papilloma virus warts. Papillomas are a benign tumor form where the skin cells are transformed by virus infection and the hypothesis was that this transformation may make them susceptible to HAMLET treatment. A population of individuals having a major problem with recurrent warts were treated topically with HAMLET and it was found that HAMLET was effective in eliminating warts and that this effect had a long lasting effect (Gustafsson NEJM 2004).

Recently, Anki Mossberg together with Bjorn Wullt performed a study on 9 patients with transitional bladder carcinomas (tumor of the urinary bladder) (Mossberg Int J Cancer 2007). They instilled HAMLET in the bladder of patients in the week preceding surgery and evaluated the tumor morphology by endoscopic photography and apoptosis-induction in biopsy tissue. They found that a reduction in tumor size was seen in 8 of 9 patients and that apoptosis was detected in the tumor tissue but not in the adjacent healthy tissue, indicating that HAMLET is specifically tumorigenic also in vivo. This study was followed up by a study in mice (Mossberg, J Urol, 2010) using the MB49 bladder carcinoma model. In this model they showed a delay in tumor development when HAMLET was instilled in the bladder.

In a study from 2014 Puthia et al (Gut January issue, 2014) it wa shown that HAMLET could be used to prevent and treat colon cancer in APC/Min-mice. HAMLET was given perorally and showed a reduced progression of tumor growth and an increased survival when given as a prophylaxis and reduction of established tumors when given as treatment. HAMLEt accumulated only in tumor tissue and only healthy cells were observed to be apoptotic.

Human milk for cancer patients?

Since the publication of the first study on HAMLET in 1995 I have been contacted by many individuals with cancer that have wondered whether drinking human milk may be beneficial for their disease. My honest answer has been that I do not know. After providing these individuals with the information I have about the potential benefits and risks of drinking milk, many individuals have gone on to try human milk as part of their therapy. There is so far no scientific study that have investigated the effect of human milk as treatment against tumor progression or the symptoms associated with tumor disease. However, most individuals that I have had contact with indicate that the milk has had some beneficial effect on their situation. Should you want more information related to these issues, please do not hesitate to contact me by e-mail.

My laboratory is not currently involved in projects directly assessing the tumor activity of HAMLET. For more information go to the Svanborg web-site.