Note that since $\det$ is multi linear and alternating, we have
\begin{eqnarray}
D(x) &=& \det \begin{bmatrix}
1 & x_1+a & x_1^2+bx_1+c \\
1 & x_2+a & x_2^2+bx_2+c \\
1 & x+a & x^2+bx+c \
\end{bmatrix} \\
&=& \det \begin{bmatrix}
1 & x_1 & x_1^2+bx_1+c \\
1 & x_2 & x_2^2+bx_2+c \\
1 & x & x^2+bx+c \
\end{bmatrix} \\
&=& \det \begin{bmatrix}
1 & x_1 & x_1^2+bx_1 \\
1 & x_2 & x_2^2+bx_2 \\
1 & x & x^2+bx \
\end{bmatrix} \\
&=& \det \begin{bmatrix}
1 & x_1 & x_1^2 \\
1 & x_2 & x_2^2 \\
1 & x & x^2 \
\end{bmatrix} \\
\end{eqnarray}
We see from this that $D$ is a second degree polynomial in $x$ and
$D(x_1) = D(x_2) = 0$. We can read off the coefficient $m$ as the coefficient of $x^2$ above, $m=x_2-x_1$.
(As Joonas notes below, the above is a Vandermonde matrix which has a well
known formula for the determinant.)