Fig 1.
Anecdotal case of regression in a clam with naturally acquired BTN.
While maintaining soft-shell clams (Mya arenaria) in laboratory conditions, we repeatedly sampled hemolymph and used a sensitive qPCR assay to determine the fraction of the cells in the hemolymph that are MarBTN. In the case of one clam (MLN-4F2, collected from the wild by a commercial source in Maine), we noticed that instead of progression to death, the clam appeared to progress to >20% MarBTN, but then the cancer regressed, and the animal continued to survive for several months. Black dot marks final hemolymph sample before death.
Fig 2.
Multiple distinct outcomes observed in soft-shell clams (Mya arenaria) after natural infection with MarBTN.
Soft-shell clams were collected from multiple sites in Maine, USA, diagnosed for MarBTN using qPCR, and we selected 21 low-positive clams (<10% MarBTN in hemolymph) and 39 negative control clams (undetectable for MarBTN-specific DNA) to follow over time, sampling hemolymph every two weeks until death. Each colored line is a separate clam. Large dots represent final hemolymph draw before clam death. Dotted line shows 10% cancer. Animals that started with low cancer have been split into three categories reflecting different outcomes: Progression (n = 9), Long-term non-progression (n = 11), and Regression (n = 1). Negative control clams (n = 39) are all shown on the bottom.
Fig 3.
Survival of soft-shell clams with MarBTN compared to control clams.
(A) A Kaplan-Meier survival curve compares naturally-infected clams with low levels of MarBTN (<10% at Day 0, n = 21, red line) with paired negative controls (MarBTN undetectable at Day 0, blue line). (B) Analysis of long-term non-progressing clams (n = 11, light red line), along with their respective paired controls (blue line), shows no difference in survival with maintained low-level infection. (C) A sub-analysis of low-positive clams in which MarBTN progressed and did not regress (n = 9, dark red line), with survival starting at the time the clam was detected with >10% cancer, compared with their paired control clams starting at the same date (blue line), shows a significant decrease in time-to death. “+” marks the dates at which two negative control animals were culled before natural death (these events were censored in the survival analysis).
Fig 4.
Detection of MarBTN-specific eDNA in tank water at late stages of MarBTN progression.
For all cases of progression (n = 9) and regression (n = 1), and a subset of cases of long-tern non-progression (n = 6) and negative controls (n = 9), eDNA was extracted from tank water collected throughout the study, and MarBTN-specific eDNA was detected using qPCR. Dual axis plots compare the amount of MarBTN cells in the hemolymph (red and blue lines, left axis) with quantity of MarBTN-specific eDNA per mL in the tank at the time of collection (black points, right axis). Representative clams are shown for (A) progression, (B) long-term non-progression, (C) regression, and (D) negative controls (see S1 Fig for comparison with microscopy and S2 Fig for plots of qPCR measurements for all individuals). A heatmap (E) of the amount of MarBTN in hemolymph over time for multiple clams of each category is shown next to a parallel heatmap (F) of detection of MarBTN-specific eDNA showing that within the progression of each animal, MarBTN DNA release occurs mostly at later stages of disease progression.
Fig 5.
Modeling the relationship between cancer level in hemolymph and MarBTN in tank eDNA.
For each timepoint for each clam at which both eDNA and hemolymph qPCR results are available, we plot a point comparing the two. Level of MarBTN in the clam is determined by the percent of MarBTN cells in hemolymph based on qPCR analysis of hemocyte genomic DNA, and cancer cell release within the 24 hrs before collection is estimated with the log transformed quantity of MarBTN-specific eDNA extracted from tank water. The vast majority of samples from clams with undetectable and very low levels of cancer in the hemolymph are associated with no MarBTN DNA release into the tank water, although there a small number of cases with detectible levels. Changepoint analysis (S2 Fig), shows that MarBTN-specific DNA is rarely detected in tank water when clams have less than 24% MarBTN in their hemolymph, so a linear regression was fit to the data points above this value (blue line, grey area marks 95% confidence interval). This shows a significant positive correlation between MarBTN-specific DNA in the water and disease progression at MarBTN levels above 24%. For the linear regression, there was an intercept of -1.2586, coefficient of 3.1869, and p-value of 0.0001258.