The Bacterial Chaperone High-Temperature Protein G (HtpG) Decreases in Abundance with Induction Therapy in Children with Acute Lymphoblastic Leukemia

Introduction: Heat shock proteins (HSPs) are ancient and highly conserved genes that participate in protein quality control in eukaryotes and bacteria. These proteins are produced in large quantities when cells are exposed to stress that causes denaturation of essential proteins. HSPs help maintain cellular homeostasis, and act as molecular chaperones, assisting with protein folding and stability, multi-protein assembly, intracellular protein trafficking and degrading damaged proteins. The Hsp90 family of proteins includes 4 genes, with 1 (HtpG) being a bacterial form. Members of the Hsp90 family interact with a diverse set of client proteins, some involved in cell signalling and proliferation. Many of these client proteins are involved in growth, survival and adaptation of cancer cells. Eukaryotic Hsp90 plays an important role in tumorigenesis and there has been much research into the use of Hsp90 inhibitors to inhibit growth of cancer cells, but little has been done to examine what role the other paralogues have in stabilizing cancer cells. In addition, no work to date has looked at whether bacterial HtpG paralogues might interact with cancer cells. While it is anticipated that bacterial stress and the behaviour of HtpG would correlate to the host response to stress, this has not been examined. The first step is to examine bacterial HtpG gene levels before and during induction to determine if there are changes in HtpG gene abundance.

Methods: Forty-eight stool samples were collected from 29 pediatric patients with Acute Lymphoblastic Leukemia (ALL) being treated at the IWK health centre, Nova Scotia, Canada after informed consent. DNA was extracted and whole shotgun metagenome sequencing (Illumina, Nextera XT) was performed. Low quality and contaminated sequences (human and PhiX174) were removed using the kneaddata pipeline (Trimmomatic and Bowtie2). Paired cleaned sequences were concatenated and HUMAnN3 was used to assign gene families to reads using the Uniref90 database. Genes assigned to HtpG or "Hsp90-like" gene families were included and summed from each sample and corrected for sequencing depth. Samples were grouped by phase of treatment (13 pre-treatment samples from 13 patients and 35 induction samples from 23 patients) for analysis. Comparisons between pre-treatment (PT) and induction (IN) samples were performed using Wilcoxon-Mann-Whitney U. Fisher's exact test was used to compare age group and sex between PT and IN samples. The study was approved by the IWK ethics board.

Results:

Neither sex nor age category differed between PT and IN (p=1). HtpG gene counts among the 48 stool samples from 29 patients ranged from 0 to 329 (mean 119). The mean counts of HtpG genes were significantly decreased in samples during induction therapy compared to pre-treatment (p=0.002). Shannon diversity, a measure of microbial diversity, did not differ between PT and IN samples (p=0.2793), nor did counts of bacteria (Chao-1 p=0.47).

Discussion: Studies have found that Hsp90 is generally over expressed in cancer patients with elevated levels, suggesting poorer overall survival. They have been proposed as a potential biomarker of poor prognosis. We had anticipated increased stress and protein damage to occur in the gut environment as a result of chemotherapy, rather than as a result of the underlying cancer. The anticipated increase in HtpG during induction did not occur but rather declined with cancer remission. One possibility is that the decline in HtpG genes was a result of a loss of bacterial diversity due to treatment. However, we did not see significant differences in diversity between PT and IN. The increase in HtpG genes during PT and reduction during induction therapy seems to follow a similar pattern to Hsp90 expression in host cells, suggesting that bacterial HtpG might also serve as a marker of disease remission. It also highlights that a better understanding of bacterial HtpG, a protein known to be present in outer membrane vesicles, which are crucial for host communication and have been implicated in contributing to carcinogenic potential is warranted.

Conclusion: This study suggests that the bacterial chaperone HtpG is possibly related to degree of inflammation caused by leukemic cells rather than leukemia treatment. A larger study is needed to verify the potential of HtpG as a marker of cancer remission or treatment toxicity or cancer immune response.

No relevant conflicts of interest to declare.